Introduction

Welcome to the documentation repository for Synapse, a Matrix homeserver implementation developed by the matrix.org core team.

Installing and using Synapse

This documentation covers topics for installation, configuration and maintenance of your Synapse process:

• Learn how to install and configure your own instance, perhaps with Single Sign-On.

• See how to upgrade between Synapse versions.

• Administer your instance using the Admin API, installing pluggable modules, or by accessing the manhole.

• Learn how to read log lines, configure logging or set up structured logging.

• Scale Synapse through additional worker processes.

• Set up monitoring and metrics to keep an eye on your Synapse instance's performance.

Developing on Synapse

Contributions are welcome! Synapse is primarily written in Python. As a developer, you may be interested in the following documentation:

• Read the Contributing Guide. It is meant to walk new contributors through the process of developing and submitting a change to the Synapse codebase (which is hosted on GitHub).

• Set up your development environment, then learn how to lint and test your code.

• Look at the issue tracker for bugs to fix or features to add. If you're new, it may be best to start with those labeled good first issue.

• Understand how Synapse is built, how to migrate database schemas, learn about federation and how to set up a local federation for development.

• We like to keep our git history clean. Learn how to do so!

• And finally, contribute to this documentation! The source for which is located here.

Donating to Synapse development

Want to help keep Synapse going but don't know how to code? Synapse is a Matrix.org Foundation project. Consider becoming a supporter on Liberapay, Patreon or through PayPal via a one-time donation.

If you are an organisation or enterprise and would like to sponsor development, reach out to us over email at: support (at) matrix.org

Reporting a security vulnerability

If you've found a security issue in Synapse or any other Matrix.org Foundation project, please report it to us in accordance with our Security Disclosure Policy. Thank you!

Installation Instructions

Choosing your server name

It is important to choose the name for your server before you install Synapse, because it cannot be changed later.

The server name determines the "domain" part of user-ids for users on your server: these will all be of the format @user:my.domain.name. It also determines how other matrix servers will reach yours for federation.

For a test configuration, set this to the hostname of your server. For a more production-ready setup, you will probably want to specify your domain (example.com) rather than a matrix-specific hostname here (in the same way that your email address is probably user@example.com rather than user@email.example.com) - but doing so may require more advanced setup: see Setting up Federation.

Installing Synapse

Prebuilt packages

Prebuilt packages are available for a number of platforms. These are recommended for most users.

Docker images and Ansible playbooks

There is an official synapse image available at https://hub.docker.com/r/matrixdotorg/synapse which can be used with the docker-compose file available at contrib/docker. Further information on this including configuration options is available in the README on hub.docker.com.

Alternatively, Andreas Peters (previously Silvio Fricke) has contributed a Dockerfile to automate a synapse server in a single Docker image, at https://hub.docker.com/r/avhost/docker-matrix/tags/

Slavi Pantaleev has created an Ansible playbook, which installs the offical Docker image of Matrix Synapse along with many other Matrix-related services (Postgres database, Element, coturn, ma1sd, SSL support, etc.). For more details, see https://github.com/spantaleev/matrix-docker-ansible-deploy

Debian/Ubuntu

Matrix.org packages

Matrix.org provides Debian/Ubuntu packages of Synapse, for the amd64 architecture via https://packages.matrix.org/debian/.

To install the latest release:

sudo apt install -y lsb-release wget apt-transport-https
sudo wget -O /usr/share/keyrings/matrix-org-archive-keyring.gpg https://packages.matrix.org/debian/matrix-org-archive-keyring.gpg
echo "deb [signed-by=/usr/share/keyrings/matrix-org-archive-keyring.gpg] https://packages.matrix.org/debian/ $(lsb_release -cs) main" |
    sudo tee /etc/apt/sources.list.d/matrix-org.list
sudo apt update
sudo apt install matrix-synapse-py3

Packages are also published for release candidates. To enable the prerelease channel, add prerelease to the sources.list line. For example:

sudo wget -O /usr/share/keyrings/matrix-org-archive-keyring.gpg https://packages.matrix.org/debian/matrix-org-archive-keyring.gpg
echo "deb [signed-by=/usr/share/keyrings/matrix-org-archive-keyring.gpg] https://packages.matrix.org/debian/ $(lsb_release -cs) main prerelease" |
    sudo tee /etc/apt/sources.list.d/matrix-org.list
sudo apt update
sudo apt install matrix-synapse-py3

The fingerprint of the repository signing key (as shown by gpg /usr/share/keyrings/matrix-org-archive-keyring.gpg) is AAF9AE843A7584B5A3E4CD2BCF45A512DE2DA058.

When installing with Debian packages, you might prefer to place files in /etc/matrix-synapse/conf.d/ to override your configuration without editing the main configuration file at /etc/matrix-synapse/homeserver.yaml. By doing that, you won't be asked if you want to replace your configuration file when you upgrade the Debian package to a later version.

Downstream Debian packages

Andrej Shadura maintains a matrix-synapse package in the Debian repositories. For bookworm and sid, it can be installed simply with:

sudo apt install matrix-synapse

Synapse is also avaliable in bullseye-backports. Please see the Debian documentation for information on how to use backports.

matrix-synapse is no longer maintained for buster and older.

Downstream Ubuntu packages

We do not recommend using the packages in the default Ubuntu repository at this time, as they are old and suffer from known security vulnerabilities. The latest version of Synapse can be installed from our repository.

Fedora

Synapse is in the Fedora repositories as matrix-synapse:

sudo dnf install matrix-synapse

Oleg Girko provides Fedora RPMs at https://obs.infoserver.lv/project/monitor/matrix-synapse

OpenSUSE

Synapse is in the OpenSUSE repositories as matrix-synapse:

sudo zypper install matrix-synapse

SUSE Linux Enterprise Server

Unofficial package are built for SLES 15 in the openSUSE:Backports:SLE-15 repository at https://download.opensuse.org/repositories/openSUSE:/Backports:/SLE-15/standard/

ArchLinux

The quickest way to get up and running with ArchLinux is probably with the community package https://www.archlinux.org/packages/community/any/matrix-synapse/, which should pull in most of the necessary dependencies.

pip may be outdated (6.0.7-1 and needs to be upgraded to 6.0.8-1 ):

sudo pip install --upgrade pip

If you encounter an error with lib bcrypt causing an Wrong ELF Class: ELFCLASS32 (x64 Systems), you may need to reinstall py-bcrypt to correctly compile it under the right architecture. (This should not be needed if installing under virtualenv):

sudo pip uninstall py-bcrypt
sudo pip install py-bcrypt

Void Linux

Synapse can be found in the void repositories as 'synapse':

xbps-install -Su
xbps-install -S synapse

FreeBSD

Synapse can be installed via FreeBSD Ports or Packages contributed by Brendan Molloy from:

• Ports: cd /usr/ports/net-im/py-matrix-synapse && make install clean
• Packages: pkg install py38-matrix-synapse

OpenBSD

As of OpenBSD 6.7 Synapse is available as a pre-compiled binary. The filesystem underlying the homeserver directory (defaults to /var/synapse) has to be mounted with wxallowed (cf. mount(8)), so creating a separate filesystem and mounting it to /var/synapse should be taken into consideration.

Installing Synapse:

doas pkg_add synapse

NixOS

Robin Lambertz has packaged Synapse for NixOS at: https://github.com/NixOS/nixpkgs/blob/master/nixos/modules/services/misc/matrix-synapse.nix

Installing as a Python module from PyPI

It's also possible to install Synapse as a Python module from PyPI.

When following this route please make sure that the Platform-specific prerequisites are already installed.

System requirements:

• POSIX-compliant system (tested on Linux & OS X)
• Python 3.7 or later, up to Python 3.10.
• At least 1GB of free RAM if you want to join large public rooms like #matrix:matrix.org

If building on an uncommon architecture for which pre-built wheels are unavailable, you will need to have a recent Rust compiler installed. The easiest way of installing the latest version is to use rustup.

To install the Synapse homeserver run:

mkdir -p ~/synapse
virtualenv -p python3 ~/synapse/env
source ~/synapse/env/bin/activate
pip install --upgrade pip
pip install --upgrade setuptools
pip install matrix-synapse

This will download Synapse from PyPI and install it, along with the python libraries it uses, into a virtual environment under ~/synapse/env. Feel free to pick a different directory if you prefer.

This Synapse installation can then be later upgraded by using pip again with the update flag:

source ~/synapse/env/bin/activate
pip install -U matrix-synapse

Before you can start Synapse, you will need to generate a configuration file. To do this, run (in your virtualenv, as before):

cd ~/synapse
python -m synapse.app.homeserver \
    --server-name my.domain.name \
    --config-path homeserver.yaml \
    --generate-config \
    --report-stats=[yes|no]

... substituting an appropriate value for --server-name and choosing whether or not to report usage statistics (hostname, Synapse version, uptime, total users, etc.) to the developers via the --report-stats argument.

This command will generate you a config file that you can then customise, but it will also generate a set of keys for you. These keys will allow your homeserver to identify itself to other homeserver, so don't lose or delete them. It would be wise to back them up somewhere safe. (If, for whatever reason, you do need to change your homeserver's keys, you may find that other homeserver have the old key cached. If you update the signing key, you should change the name of the key in the <server name>.signing.key file (the second word) to something different. See the spec for more information on key management).

To actually run your new homeserver, pick a working directory for Synapse to run (e.g. ~/synapse), and:

cd ~/synapse
source env/bin/activate
synctl start

Platform-specific prerequisites

Synapse is written in Python but some of the libraries it uses are written in C. So before we can install Synapse itself we need a working C compiler and the header files for Python C extensions.

Debian/Ubuntu/Raspbian

Installing prerequisites on Ubuntu or Debian:

sudo apt install build-essential python3-dev libffi-dev \
                     python3-pip python3-setuptools sqlite3 \
                     libssl-dev virtualenv libjpeg-dev libxslt1-dev

ArchLinux

Installing prerequisites on ArchLinux:

sudo pacman -S base-devel python python-pip \
               python-setuptools python-virtualenv sqlite3

CentOS/Fedora

Installing prerequisites on CentOS or Fedora Linux:

sudo dnf install libtiff-devel libjpeg-devel libzip-devel freetype-devel \
                 libwebp-devel libxml2-devel libxslt-devel libpq-devel \
                 python3-virtualenv libffi-devel openssl-devel python3-devel
sudo dnf groupinstall "Development Tools"

macOS

Installing prerequisites on macOS:

You may need to install the latest Xcode developer tools:

xcode-select --install

On ARM-based Macs you may need to explicitly install libjpeg which is a pillow dependency. You can use Homebrew (https://brew.sh):

 brew install jpeg

On macOS Catalina (10.15) you may need to explicitly install OpenSSL via brew and inform pip about it so that psycopg2 builds:

brew install openssl@1.1
export LDFLAGS="-L/usr/local/opt/openssl/lib"
export CPPFLAGS="-I/usr/local/opt/openssl/include"

OpenSUSE

Installing prerequisites on openSUSE:

sudo zypper in -t pattern devel_basis
sudo zypper in python-pip python-setuptools sqlite3 python-virtualenv \
               python-devel libffi-devel libopenssl-devel libjpeg62-devel

OpenBSD

A port of Synapse is available under net/synapse. The filesystem underlying the homeserver directory (defaults to /var/synapse) has to be mounted with wxallowed (cf. mount(8)), so creating a separate filesystem and mounting it to /var/synapse should be taken into consideration.

To be able to build Synapse's dependency on python the WRKOBJDIR (cf. bsd.port.mk(5)) for building python, too, needs to be on a filesystem mounted with wxallowed (cf. mount(8)).

Creating a WRKOBJDIR for building python under /usr/local (which on a default OpenBSD installation is mounted with wxallowed):

doas mkdir /usr/local/pobj_wxallowed

Assuming PORTS_PRIVSEP=Yes (cf. bsd.port.mk(5)) and SUDO=doas are configured in /etc/mk.conf:

doas chown _pbuild:_pbuild /usr/local/pobj_wxallowed

Setting the WRKOBJDIR for building python:

echo WRKOBJDIR_lang/python/3.7=/usr/local/pobj_wxallowed  \\nWRKOBJDIR_lang/python/2.7=/usr/local/pobj_wxallowed >> /etc/mk.conf

Building Synapse:

cd /usr/ports/net/synapse
make install

Windows

Running Synapse natively on Windows is not officially supported.

If you wish to run or develop Synapse on Windows, the Windows Subsystem for Linux provides a Linux environment which is capable of using the Debian, Fedora, or source installation methods. More information about WSL can be found at https://docs.microsoft.com/en-us/windows/wsl/install for Windows 10/11 and https://docs.microsoft.com/en-us/windows/wsl/install-on-server for Windows Server.

Setting up Synapse

Once you have installed synapse as above, you will need to configure it.

Using PostgreSQL

By default Synapse uses an SQLite database and in doing so trades performance for convenience. Almost all installations should opt to use PostgreSQL instead. Advantages include:

• significant performance improvements due to the superior threading and caching model, smarter query optimiser
• allowing the DB to be run on separate hardware

For information on how to install and use PostgreSQL in Synapse, please see Using Postgres

SQLite is only acceptable for testing purposes. SQLite should not be used in a production server. Synapse will perform poorly when using SQLite, especially when participating in large rooms.

TLS certificates

The default configuration exposes a single HTTP port on the local interface: http://localhost:8008. It is suitable for local testing, but for any practical use, you will need Synapse's APIs to be served over HTTPS.

The recommended way to do so is to set up a reverse proxy on port 8448. You can find documentation on doing so in the reverse proxy documentation.

Alternatively, you can configure Synapse to expose an HTTPS port. To do so, you will need to edit homeserver.yaml, as follows:

• First, under the listeners option, add the configuration for the TLS-enabled listener like so:

listeners:
  - port: 8448
    type: http
    tls: true
    resources:
      - names: [client, federation]

• You will also need to add the options tls_certificate_path and tls_private_key_path. to your configuration file. You will need to manage provisioning of these certificates yourself.

• You can find more information about these options as well as how to configure synapse in the configuration manual.

  If you are using your own certificate, be sure to use a .pem file that includes the full certificate chain including any intermediate certificates (for instance, if using certbot, use fullchain.pem as your certificate, not cert.pem).

For a more detailed guide to configuring your server for federation, see Federation.

Client Well-Known URI

Setting up the client Well-Known URI is optional but if you set it up, it will allow users to enter their full username (e.g. @user:<server_name>) into clients which support well-known lookup to automatically configure the homeserver and identity server URLs. This is useful so that users don't have to memorize or think about the actual homeserver URL you are using.

The URL https://<server_name>/.well-known/matrix/client should return JSON in the following format.

{
  "m.homeserver": {
    "base_url": "https://<matrix.example.com>"
  }
}

It can optionally contain identity server information as well.

{
  "m.homeserver": {
    "base_url": "https://<matrix.example.com>"
  },
  "m.identity_server": {
    "base_url": "https://<identity.example.com>"
  }
}

To work in browser based clients, the file must be served with the appropriate Cross-Origin Resource Sharing (CORS) headers. A recommended value would be Access-Control-Allow-Origin: * which would allow all browser based clients to view it.

In nginx this would be something like:

location /.well-known/matrix/client {
    return 200 '{"m.homeserver": {"base_url": "https://<matrix.example.com>"}}';
    default_type application/json;
    add_header Access-Control-Allow-Origin *;
}

You should also ensure the public_baseurl option in homeserver.yaml is set correctly. public_baseurl should be set to the URL that clients will use to connect to your server. This is the same URL you put for the m.homeserver base_url above.

public_baseurl: "https://<matrix.example.com>"

Email

It is desirable for Synapse to have the capability to send email. This allows Synapse to send password reset emails, send verifications when an email address is added to a user's account, and send email notifications to users when they receive new messages.

To configure an SMTP server for Synapse, modify the configuration section headed email, and be sure to have at least the smtp_host, smtp_port and notif_from fields filled out. You may also need to set smtp_user, smtp_pass, and require_transport_security.

If email is not configured, password reset, registration and notifications via email will be disabled.

Registering a user

One way to create a new user is to do so from a client like Element. This requires registration to be enabled via the enable_registration setting.

Alternatively, you can create new users from the command line. This can be done as follows:

1. If synapse was installed via pip, activate the virtualenv as follows (if Synapse was installed via a prebuilt package, register_new_matrix_user should already be on the search path):

   cd ~/synapse
   source env/bin/activate
   synctl start # if not already running
   

2. Run the following command:

   register_new_matrix_user -c homeserver.yaml
   

This will prompt you to add details for the new user, and will then connect to the running Synapse to create the new user. For example:

New user localpart: erikj
Password:
Confirm password:
Make admin [no]:
Success!

This process uses a setting registration_shared_secret, which is shared between Synapse itself and the register_new_matrix_user script. It doesn't matter what it is (a random value is generated by --generate-config), but it should be kept secret, as anyone with knowledge of it can register users, including admin accounts, on your server even if enable_registration is false.

Setting up a TURN server

For reliable VoIP calls to be routed via this homeserver, you MUST configure a TURN server. See TURN setup for details.

URL previews

Synapse includes support for previewing URLs, which is disabled by default. To turn it on you must enable the url_preview_enabled: True config parameter and explicitly specify the IP ranges that Synapse is not allowed to spider for previewing in the url_preview_ip_range_blacklist configuration parameter. This is critical from a security perspective to stop arbitrary Matrix users spidering 'internal' URLs on your network. At the very least we recommend that your loopback and RFC1918 IP addresses are blacklisted.

This also requires the optional lxml python dependency to be installed. This in turn requires the libxml2 library to be available - on Debian/Ubuntu this means apt-get install libxml2-dev, or equivalent for your OS.

Troubleshooting Installation

pip seems to leak lots of memory during installation. For instance, a Linux host with 512MB of RAM may run out of memory whilst installing Twisted. If this happens, you will have to individually install the dependencies which are failing, e.g.:

pip install twisted

If you have any other problems, feel free to ask in #synapse:matrix.org.

Using Postgres

Synapse supports PostgreSQL versions 10 or later.

Install postgres client libraries

Synapse will require the python postgres client library in order to connect to a postgres database.

• If you are using the matrix.org debian/ubuntu packages, the necessary python library will already be installed, but you will need to ensure the low-level postgres library is installed, which you can do with apt install libpq5.

• For other pre-built packages, please consult the documentation from the relevant package.

• If you installed synapse in a virtualenv, you can install the library with:

  ~/synapse/env/bin/pip install "matrix-synapse[postgres]"
  

  (substituting the path to your virtualenv for ~/synapse/env, if you used a different path). You will require the postgres development files. These are in the libpq-dev package on Debian-derived distributions.

Set up database

Assuming your PostgreSQL database user is called postgres, first authenticate as the database user with:

su - postgres
# Or, if your system uses sudo to get administrative rights
sudo -u postgres bash

Then, create a postgres user and a database with:

# this will prompt for a password for the new user
createuser --pwprompt synapse_user

createdb --encoding=UTF8 --locale=C --template=template0 --owner=synapse_user synapse

The above will create a user called synapse_user, and a database called synapse.

Note that the PostgreSQL database must have the correct encoding set (as shown above), otherwise it will not be able to store UTF8 strings.

You may need to enable password authentication so synapse_user can connect to the database. See https://www.postgresql.org/docs/current/auth-pg-hba-conf.html.

Synapse config

When you are ready to start using PostgreSQL, edit the database section in your config file to match the following lines:

database:
  name: psycopg2
  args:
    user: <user>
    password: <pass>
    database: <db>
    host: <host>
    cp_min: 5
    cp_max: 10

All key, values in args are passed to the psycopg2.connect(..) function, except keys beginning with cp_, which are consumed by the twisted adbapi connection pool. See the libpq documentation for a list of options which can be passed.

You should consider tuning the args.keepalives_* options if there is any danger of the connection between your homeserver and database dropping, otherwise Synapse may block for an extended period while it waits for a response from the database server. Example values might be:

database:
  args:
    # ... as above

    # seconds of inactivity after which TCP should send a keepalive message to the server
    keepalives_idle: 10

    # the number of seconds after which a TCP keepalive message that is not
    # acknowledged by the server should be retransmitted
    keepalives_interval: 10

    # the number of TCP keepalives that can be lost before the client's connection
    # to the server is considered dead
    keepalives_count: 3

Tuning Postgres

The default settings should be fine for most deployments. For larger scale deployments tuning some of the settings is recommended, details of which can be found at https://wiki.postgresql.org/wiki/Tuning_Your_PostgreSQL_Server.

In particular, we've found tuning the following values helpful for performance:

• shared_buffers
• effective_cache_size
• work_mem
• maintenance_work_mem
• autovacuum_work_mem

Note that the appropriate values for those fields depend on the amount of free memory the database host has available.

Additionally, admins of large deployments might want to consider using huge pages to help manage memory, especially when using large values of shared_buffers. You can read more about that here.

Porting from SQLite

Overview

The script synapse_port_db allows porting an existing synapse server backed by SQLite to using PostgreSQL. This is done in as a two phase process:

1. Copy the existing SQLite database to a separate location and run the port script against that offline database.
2. Shut down the server. Rerun the port script to port any data that has come in since taking the first snapshot. Restart server against the PostgreSQL database.

The port script is designed to be run repeatedly against newer snapshots of the SQLite database file. This makes it safe to repeat step 1 if there was a delay between taking the previous snapshot and being ready to do step 2.

It is safe to at any time kill the port script and restart it.

However, under no circumstances should the SQLite database be VACUUMed between multiple runs of the script. Doing so can lead to an inconsistent copy of your database into Postgres. To avoid accidental error, the script will check that SQLite's auto_vacuum mechanism is disabled, but the script is not able to protect against a manual VACUUM operation performed either by the administrator or by any automated task that the administrator may have configured.

Note that the database may take up significantly more (25% - 100% more) space on disk after porting to Postgres.

Using the port script

Firstly, shut down the currently running synapse server and copy its database file (typically homeserver.db) to another location. Once the copy is complete, restart synapse. For instance:

synctl stop
cp homeserver.db homeserver.db.snapshot
synctl start

Copy the old config file into a new config file:

cp homeserver.yaml homeserver-postgres.yaml

Edit the database section as described in the section Synapse config above and with the SQLite snapshot located at homeserver.db.snapshot simply run:

synapse_port_db --sqlite-database homeserver.db.snapshot \
    --postgres-config homeserver-postgres.yaml

The flag --curses displays a coloured curses progress UI.

If the script took a long time to complete, or time has otherwise passed since the original snapshot was taken, repeat the previous steps with a newer snapshot.

To complete the conversion shut down the synapse server and run the port script one last time, e.g. if the SQLite database is at homeserver.db run:

synapse_port_db --sqlite-database homeserver.db \
    --postgres-config homeserver-postgres.yaml

Once that has completed, change the synapse config to point at the PostgreSQL database configuration file homeserver-postgres.yaml:

synctl stop
mv homeserver.yaml homeserver-old-sqlite.yaml
mv homeserver-postgres.yaml homeserver.yaml
synctl start

Synapse should now be running against PostgreSQL.

Troubleshooting

Alternative auth methods

If you get an error along the lines of FATAL: Ident authentication failed for user "synapse_user", you may need to use an authentication method other than ident:

• If the synapse_user user has a password, add the password to the database: section of homeserver.yaml. Then add the following to pg_hba.conf:

  host    synapse     synapse_user    ::1/128     md5  # or `scram-sha-256` instead of `md5` if you use that
  

• If the synapse_user user does not have a password, then a password doesn't have to be added to homeserver.yaml. But the following does need to be added to pg_hba.conf:

  host    synapse     synapse_user    ::1/128     trust
  

Note that line order matters in pg_hba.conf, so make sure that if you do add a new line, it is inserted before:

host    all         all             ::1/128     ident

Fixing incorrect COLLATE or CTYPE

Synapse will refuse to set up a new database if it has the wrong values of COLLATE and CTYPE set. Synapse will also refuse to start an existing database with incorrect values of COLLATE and CTYPE unless the config flag allow_unsafe_locale, found in the database section of the config, is set to true. Using different locales can cause issues if the locale library is updated from underneath the database, or if a different version of the locale is used on any replicas.

If you have a databse with an unsafe locale, the safest way to fix the issue is to dump the database and recreate it with the correct locale parameter (as shown above). It is also possible to change the parameters on a live database and run a REINDEX on the entire database, however extreme care must be taken to avoid database corruption.

Note that the above may fail with an error about duplicate rows if corruption has already occurred, and such duplicate rows will need to be manually removed.

Fixing inconsistent sequences error

Synapse uses Postgres sequences to generate IDs for various tables. A sequence and associated table can get out of sync if, for example, Synapse has been downgraded and then upgraded again.

To fix the issue shut down Synapse (including any and all workers) and run the SQL command included in the error message. Once done Synapse should start successfully.

Using a reverse proxy with Synapse

It is recommended to put a reverse proxy such as nginx, Apache, Caddy, HAProxy or relayd in front of Synapse. One advantage of doing so is that it means that you can expose the default https port (443) to Matrix clients without needing to run Synapse with root privileges.

You should configure your reverse proxy to forward requests to /_matrix or /_synapse/client to Synapse, and have it set the X-Forwarded-For and X-Forwarded-Proto request headers.

You should remember that Matrix clients and other Matrix servers do not necessarily need to connect to your server via the same server name or port. Indeed, clients will use port 443 by default, whereas servers default to port 8448. Where these are different, we refer to the 'client port' and the 'federation port'. See the Matrix specification for more details of the algorithm used for federation connections, and Delegation for instructions on setting up delegation.

NOTE: Your reverse proxy must not canonicalise or normalise the requested URI in any way (for example, by decoding %xx escapes). Beware that Apache will canonicalise URIs unless you specify nocanon.

Let's assume that we expect clients to connect to our server at https://matrix.example.com, and other servers to connect at https://example.com:8448. The following sections detail the configuration of the reverse proxy and the homeserver.

Homeserver Configuration

The HTTP configuration will need to be updated for Synapse to correctly record client IP addresses and generate redirect URLs while behind a reverse proxy.

In homeserver.yaml set x_forwarded: true in the port 8008 section and consider setting bind_addresses: ['127.0.0.1'] so that the server only listens to traffic on localhost. (Do not change bind_addresses to 127.0.0.1 when using a containerized Synapse, as that will prevent it from responding to proxied traffic.)

Reverse-proxy configuration examples

NOTE: You only need one of these.

nginx

server {
    listen 443 ssl http2;
    listen [::]:443 ssl http2;

    # For the federation port
    listen 8448 ssl http2 default_server;
    listen [::]:8448 ssl http2 default_server;

    server_name matrix.example.com;

    location ~ ^(/_matrix|/_synapse/client) {
        # note: do not add a path (even a single /) after the port in `proxy_pass`,
        # otherwise nginx will canonicalise the URI and cause signature verification
        # errors.
        proxy_pass http://localhost:8008;
        proxy_set_header X-Forwarded-For $remote_addr;
        proxy_set_header X-Forwarded-Proto $scheme;
        proxy_set_header Host $host;

        # Nginx by default only allows file uploads up to 1M in size
        # Increase client_max_body_size to match max_upload_size defined in homeserver.yaml
        client_max_body_size 50M;
    }
}

Caddy v2

matrix.example.com {
  reverse_proxy /_matrix/* localhost:8008
  reverse_proxy /_synapse/client/* localhost:8008
}

example.com:8448 {
  reverse_proxy localhost:8008
}

Delegation example:

example.com {
	header /.well-known/matrix/* Content-Type application/json
	header /.well-known/matrix/* Access-Control-Allow-Origin *
	respond /.well-known/matrix/server `{"m.server": "matrix.example.com:443"}`
	respond /.well-known/matrix/client `{"m.homeserver":{"base_url":"https://matrix.example.com"},"m.identity_server":{"base_url":"https://identity.example.com"}}`
}

matrix.example.com {
    reverse_proxy /_matrix/* localhost:8008
    reverse_proxy /_synapse/client/* localhost:8008
}

Apache

<VirtualHost *:443>
    SSLEngine on
    ServerName matrix.example.com

    RequestHeader set "X-Forwarded-Proto" expr=%{REQUEST_SCHEME}
    AllowEncodedSlashes NoDecode
    ProxyPreserveHost on
    ProxyPass /_matrix http://127.0.0.1:8008/_matrix nocanon
    ProxyPassReverse /_matrix http://127.0.0.1:8008/_matrix
    ProxyPass /_synapse/client http://127.0.0.1:8008/_synapse/client nocanon
    ProxyPassReverse /_synapse/client http://127.0.0.1:8008/_synapse/client
</VirtualHost>

<VirtualHost *:8448>
    SSLEngine on
    ServerName example.com

    RequestHeader set "X-Forwarded-Proto" expr=%{REQUEST_SCHEME}
    AllowEncodedSlashes NoDecode
    ProxyPass /_matrix http://127.0.0.1:8008/_matrix nocanon
    ProxyPassReverse /_matrix http://127.0.0.1:8008/_matrix
</VirtualHost>

NOTE: ensure the nocanon options are included.

NOTE 2: It appears that Synapse is currently incompatible with the ModSecurity module for Apache (mod_security2). If you need it enabled for other services on your web server, you can disable it for Synapse's two VirtualHosts by including the following lines before each of the two </VirtualHost> above:

<IfModule security2_module>
    SecRuleEngine off
</IfModule>

NOTE 3: Missing ProxyPreserveHost on can lead to a redirect loop.

HAProxy

frontend https
  bind *:443,[::]:443 ssl crt /etc/ssl/haproxy/ strict-sni alpn h2,http/1.1
  http-request set-header X-Forwarded-Proto https if { ssl_fc }
  http-request set-header X-Forwarded-Proto http if !{ ssl_fc }
  http-request set-header X-Forwarded-For %[src]

  # Matrix client traffic
  acl matrix-host hdr(host) -i matrix.example.com matrix.example.com:443
  acl matrix-path path_beg /_matrix
  acl matrix-path path_beg /_synapse/client

  use_backend matrix if matrix-host matrix-path

frontend matrix-federation
  bind *:8448,[::]:8448 ssl crt /etc/ssl/haproxy/synapse.pem alpn h2,http/1.1
  http-request set-header X-Forwarded-Proto https if { ssl_fc }
  http-request set-header X-Forwarded-Proto http if !{ ssl_fc }
  http-request set-header X-Forwarded-For %[src]

  default_backend matrix

backend matrix
  server matrix 127.0.0.1:8008

Delegation example:

frontend https
  acl matrix-well-known-client-path path /.well-known/matrix/client
  acl matrix-well-known-server-path path /.well-known/matrix/server
  use_backend matrix-well-known-client if matrix-well-known-client-path
  use_backend matrix-well-known-server if matrix-well-known-server-path
 
backend matrix-well-known-client
  http-after-response set-header Access-Control-Allow-Origin "*"
  http-after-response set-header Access-Control-Allow-Methods "GET, POST, PUT, DELETE, OPTIONS"
  http-after-response set-header Access-Control-Allow-Headers "Origin, X-Requested-With, Content-Type, Accept, Authorization"
  http-request return status 200 content-type application/json string '{"m.homeserver":{"base_url":"https://matrix.example.com"},"m.identity_server":{"base_url":"https://identity.example.com"}}'

backend matrix-well-known-server
  http-after-response set-header Access-Control-Allow-Origin "*"
  http-after-response set-header Access-Control-Allow-Methods "GET, POST, PUT, DELETE, OPTIONS"
  http-after-response set-header Access-Control-Allow-Headers "Origin, X-Requested-With, Content-Type, Accept, Authorization"
  http-request return status 200 content-type application/json string '{"m.server":"matrix.example.com:443"}'

Relayd

table <webserver>    { 127.0.0.1 }
table <matrixserver> { 127.0.0.1 }

http protocol "https" {
    tls { no tlsv1.0, ciphers "HIGH" }
    tls keypair "example.com"
    match header set "X-Forwarded-For"   value "$REMOTE_ADDR"
    match header set "X-Forwarded-Proto" value "https"

    # set CORS header for .well-known/matrix/server, .well-known/matrix/client
    # httpd does not support setting headers, so do it here
    match request path "/.well-known/matrix/*" tag "matrix-cors"
    match response tagged "matrix-cors" header set "Access-Control-Allow-Origin" value "*"

    pass quick path "/_matrix/*"         forward to <matrixserver>
    pass quick path "/_synapse/client/*" forward to <matrixserver>

    # pass on non-matrix traffic to webserver
    pass                                 forward to <webserver>
}

relay "https_traffic" {
    listen on egress port 443 tls
    protocol "https"
    forward to <matrixserver> port 8008 check tcp
    forward to <webserver>    port 8080 check tcp
}

http protocol "matrix" {
    tls { no tlsv1.0, ciphers "HIGH" }
    tls keypair "example.com"
    block
    pass quick path "/_matrix/*"         forward to <matrixserver>
    pass quick path "/_synapse/client/*" forward to <matrixserver>
}

relay "matrix_federation" {
    listen on egress port 8448 tls
    protocol "matrix"
    forward to <matrixserver> port 8008 check tcp
}

Health check endpoint

Synapse exposes a health check endpoint for use by reverse proxies. Each configured HTTP listener has a /health endpoint which always returns 200 OK (and doesn't get logged).

Synapse administration endpoints

Endpoints for administering your Synapse instance are placed under /_synapse/admin. These require authentication through an access token of an admin user. However as access to these endpoints grants the caller a lot of power, we do not recommend exposing them to the public internet without good reason.

Using a forward proxy with Synapse

You can use Synapse with a forward or outbound proxy. An example of when this is necessary is in corporate environments behind a DMZ (demilitarized zone). Synapse supports routing outbound HTTP(S) requests via a proxy. Only HTTP(S) proxy is supported, not SOCKS proxy or anything else.

Configure

The http_proxy, https_proxy, no_proxy environment variables are used to specify proxy settings. The environment variable is not case sensitive.

• http_proxy: Proxy server to use for HTTP requests.
• https_proxy: Proxy server to use for HTTPS requests.
• no_proxy: Comma-separated list of hosts, IP addresses, or IP ranges in CIDR format which should not use the proxy. Synapse will directly connect to these hosts.

The http_proxy and https_proxy environment variables have the form: [scheme://][<username>:<password>@]<host>[:<port>]

• Supported schemes are http:// and https://. The default scheme is http:// for compatibility reasons; it is recommended to set a scheme. If scheme is set to https:// the connection uses TLS between Synapse and the proxy.

  NOTE: Synapse validates the certificates. If the certificate is not valid, then the connection is dropped.

• Default port if not given is 1080.

• Username and password are optional and will be used to authenticate against the proxy.

Examples

• HTTP_PROXY=http://USERNAME:PASSWORD@10.0.1.1:8080/
• HTTPS_PROXY=http://USERNAME:PASSWORD@proxy.example.com:8080/
• NO_PROXY=master.hostname.example.com,10.1.0.0/16,172.30.0.0/16

NOTE: Synapse does not apply the IP blacklist to connections through the proxy (since the DNS resolution is done by the proxy). It is expected that the proxy or firewall will apply blacklisting of IP addresses.

Connection types

The proxy will be used for:

• push
• url previews
• phone-home stats
• recaptcha validation
• CAS auth validation
• OpenID Connect
• Outbound federation
• Federation (checking public key revocation)
• Fetching public keys of other servers
• Downloading remote media

It will not be used for:

• Application Services
• Identity servers
• In worker configurations
  • connections between workers
  • connections from workers to Redis

Troubleshooting

If a proxy server is used with TLS (HTTPS) and no connections are established, it is most likely due to the proxy's certificates. To test this, the validation in Synapse can be deactivated.

NOTE: This has an impact on security and is for testing purposes only!

To deactivate the certificate validation, the following setting must be added to your homserver.yaml.

use_insecure_ssl_client_just_for_testing_do_not_use: true

Overview

This document explains how to enable VoIP relaying on your homeserver with TURN.

The synapse Matrix homeserver supports integration with TURN server via the TURN server REST API. This allows the homeserver to generate credentials that are valid for use on the TURN server through the use of a secret shared between the homeserver and the TURN server.

The following sections describe how to install coturn (which implements the TURN REST API) and integrate it with synapse.

Requirements

For TURN relaying with coturn to work, it must be hosted on a server/endpoint with a public IP.

Hosting TURN behind NAT requires port forwaring and for the NAT gateway to have a public IP. However, even with appropriate configuration, NAT is known to cause issues and to often not work.

coturn setup

Initial installation

The TURN daemon coturn is available from a variety of sources such as native package managers, or installation from source.

Debian installation

Just install the debian package:

apt install coturn

This will install and start a systemd service called coturn.

Source installation

1. Download the latest release from github. Unpack it and cd into the directory.

2. Configure it:

   ./configure
   

   You may need to install libevent2: if so, you should do so in the way recommended by your operating system. You can ignore warnings about lack of database support: a database is unnecessary for this purpose.

3. Build and install it:

   make
   make install
   

Configuration

1. Create or edit the config file in /etc/turnserver.conf. The relevant lines, with example values, are:

   use-auth-secret
   static-auth-secret=[your secret key here]
   realm=turn.myserver.org
   

   See turnserver.conf for explanations of the options. One way to generate the static-auth-secret is with pwgen:

   pwgen -s 64 1
   

   A realm must be specified, but its value is somewhat arbitrary. (It is sent to clients as part of the authentication flow.) It is conventional to set it to be your server name.

2. You will most likely want to configure coturn to write logs somewhere. The easiest way is normally to send them to the syslog:

   syslog
   

   (in which case, the logs will be available via journalctl -u coturn on a systemd system). Alternatively, coturn can be configured to write to a logfile - check the example config file supplied with coturn.

3. Consider your security settings. TURN lets users request a relay which will connect to arbitrary IP addresses and ports. The following configuration is suggested as a minimum starting point:

   # VoIP traffic is all UDP. There is no reason to let users connect to arbitrary TCP endpoints via the relay.
   no-tcp-relay
   
   # don't let the relay ever try to connect to private IP address ranges within your network (if any)
   # given the turn server is likely behind your firewall, remember to include any privileged public IPs too.
   denied-peer-ip=10.0.0.0-10.255.255.255
   denied-peer-ip=192.168.0.0-192.168.255.255
   denied-peer-ip=172.16.0.0-172.31.255.255
   
   # recommended additional local peers to block, to mitigate external access to internal services.
   # https://www.rtcsec.com/article/slack-webrtc-turn-compromise-and-bug-bounty/#how-to-fix-an-open-turn-relay-to-address-this-vulnerability
   no-multicast-peers
   denied-peer-ip=0.0.0.0-0.255.255.255
   denied-peer-ip=100.64.0.0-100.127.255.255
   denied-peer-ip=127.0.0.0-127.255.255.255
   denied-peer-ip=169.254.0.0-169.254.255.255
   denied-peer-ip=192.0.0.0-192.0.0.255
   denied-peer-ip=192.0.2.0-192.0.2.255
   denied-peer-ip=192.88.99.0-192.88.99.255
   denied-peer-ip=198.18.0.0-198.19.255.255
   denied-peer-ip=198.51.100.0-198.51.100.255
   denied-peer-ip=203.0.113.0-203.0.113.255
   denied-peer-ip=240.0.0.0-255.255.255.255
   
   # special case the turn server itself so that client->TURN->TURN->client flows work
   # this should be one of the turn server's listening IPs
   allowed-peer-ip=10.0.0.1
   
   # consider whether you want to limit the quota of relayed streams per user (or total) to avoid risk of DoS.
   user-quota=12 # 4 streams per video call, so 12 streams = 3 simultaneous relayed calls per user.
   total-quota=1200
   

4. Also consider supporting TLS/DTLS. To do this, add the following settings to turnserver.conf:

   # TLS certificates, including intermediate certs.
   # For Let's Encrypt certificates, use `fullchain.pem` here.
   cert=/path/to/fullchain.pem
   
   # TLS private key file
   pkey=/path/to/privkey.pem
   
   # Ensure the configuration lines that disable TLS/DTLS are commented-out or removed
   #no-tls
   #no-dtls
   

   In this case, replace the turn: schemes in the turn_uris settings below with turns:.

   We recommend that you only try to set up TLS/DTLS once you have set up a basic installation and got it working.

   NB: If your TLS certificate was provided by Let's Encrypt, TLS/DTLS will not work with any Matrix client that uses Chromium's WebRTC library. This currently includes Element Android & iOS; for more details, see their respective issues as well as the underlying WebRTC issue. Consider using a ZeroSSL certificate for your TURN server as a working alternative.

5. Ensure your firewall allows traffic into the TURN server on the ports you've configured it to listen on (By default: 3478 and 5349 for TURN traffic (remember to allow both TCP and UDP traffic), and ports 49152-65535 for the UDP relay.)

6. If your TURN server is behind NAT, the NAT gateway must have an external, publicly-reachable IP address. You must configure coturn to advertise that address to connecting clients:

   external-ip=EXTERNAL_NAT_IPv4_ADDRESS
   

   You may optionally limit the TURN server to listen only on the local address that is mapped by NAT to the external address:

   listening-ip=INTERNAL_TURNSERVER_IPv4_ADDRESS
   

   If your NAT gateway is reachable over both IPv4 and IPv6, you may configure coturn to advertise each available address:

   external-ip=EXTERNAL_NAT_IPv4_ADDRESS
   external-ip=EXTERNAL_NAT_IPv6_ADDRESS
   

   When advertising an external IPv6 address, ensure that the firewall and network settings of the system running your TURN server are configured to accept IPv6 traffic, and that the TURN server is listening on the local IPv6 address that is mapped by NAT to the external IPv6 address.

7. (Re)start the turn server:

   • If you used the Debian package (or have set up a systemd unit yourself):

     systemctl restart coturn
     

   • If you installed from source:

     bin/turnserver -o
     

Synapse setup

Your homeserver configuration file needs the following extra keys:

1. "turn_uris": This needs to be a yaml list of public-facing URIs for your TURN server to be given out to your clients. Add separate entries for each transport your TURN server supports.
2. "turn_shared_secret": This is the secret shared between your homeserver and your TURN server, so you should set it to the same string you used in turnserver.conf.
3. "turn_user_lifetime": This is the amount of time credentials generated by your homeserver are valid for (in milliseconds). Shorter times offer less potential for abuse at the expense of increased traffic between web clients and your homeserver to refresh credentials. The TURN REST API specification recommends one day (86400000).
4. "turn_allow_guests": Whether to allow guest users to use the TURN server. This is enabled by default, as otherwise VoIP will not work reliably for guests. However, it does introduce a security risk as it lets guests connect to arbitrary endpoints without having gone through a CAPTCHA or similar to register a real account.

As an example, here is the relevant section of the config file for matrix.org. The turn_uris are appropriate for TURN servers listening on the default ports, with no TLS.

turn_uris: [ "turn:turn.matrix.org?transport=udp", "turn:turn.matrix.org?transport=tcp" ]
turn_shared_secret: "n0t4ctuAllymatr1Xd0TorgSshar3d5ecret4obvIousreAsons"
turn_user_lifetime: 86400000
turn_allow_guests: True

After updating the homeserver configuration, you must restart synapse:

• If you use synctl:

  # Depending on how Synapse is installed, synctl may already be on
  # your PATH. If not, you may need to activate a virtual environment.
  synctl restart
  

• If you use systemd:

  systemctl restart matrix-synapse.service
  

... and then reload any clients (or wait an hour for them to refresh their settings).

Troubleshooting

The normal symptoms of a misconfigured TURN server are that calls between devices on different networks ring, but get stuck at "call connecting". Unfortunately, troubleshooting this can be tricky.

Here are a few things to try:

• Check that you have opened your firewall to allow TCP and UDP traffic to the TURN ports (normally 3478 and 5349).

• Check that you have opened your firewall to allow UDP traffic to the UDP relay ports (49152-65535 by default).

• Try disabling coturn's TLS/DTLS listeners and enable only its (unencrypted) TCP/UDP listeners. (This will only leave signaling traffic unencrypted; voice & video WebRTC traffic is always encrypted.)

• Some WebRTC implementations (notably, that of Google Chrome) appear to get confused by TURN servers which are reachable over IPv6 (this appears to be an unexpected side-effect of its handling of multiple IP addresses as defined by draft-ietf-rtcweb-ip-handling).

  Try removing any AAAA records for your TURN server, so that it is only reachable over IPv4.

• If your TURN server is behind NAT:

  • double-check that your NAT gateway is correctly forwarding all TURN ports (normally 3478 & 5349 for TCP & UDP TURN traffic, and 49152-65535 for the UDP relay) to the NAT-internal address of your TURN server. If advertising both IPv4 and IPv6 external addresses via the external-ip option, ensure that the NAT is forwarding both IPv4 and IPv6 traffic to the IPv4 and IPv6 internal addresses of your TURN server. When in doubt, remove AAAA records for your TURN server and specify only an IPv4 address as your external-ip.

  • ensure that your TURN server uses the NAT gateway as its default route.

• Enable more verbose logging in coturn via the verbose setting:

  verbose
  

  ... and then see if there are any clues in its logs.

• If you are using a browser-based client under Chrome, check chrome://webrtc-internals/ for insights into the internals of the negotiation. On Firefox, check the "Connection Log" on about:webrtc.

  (Understanding the output is beyond the scope of this document!)

• You can test your Matrix homeserver TURN setup with https://test.voip.librepush.net/. Note that this test is not fully reliable yet, so don't be discouraged if the test fails. Here is the github repo of the source of the tester, where you can file bug reports.

• There is a WebRTC test tool at https://webrtc.github.io/samples/src/content/peerconnection/trickle-ice/. To use it, you will need a username/password for your TURN server. You can either:

  • look for the GET /_matrix/client/r0/voip/turnServer request made by a matrix client to your homeserver in your browser's network inspector. In the response you should see username and password. Or:

  • Use the following shell commands:

    secret=staticAuthSecretHere
    
    u=$((`date +%s` + 3600)):test
    p=$(echo -n $u | openssl dgst -hmac $secret -sha1 -binary | base64)
    echo -e "username: $u\npassword: $p"
    

    Or:

  • Temporarily configure coturn to accept a static username/password. To do this, comment out use-auth-secret and static-auth-secret and add the following:

    lt-cred-mech
    user=username:password
    

    Note: these settings will not take effect unless use-auth-secret and static-auth-secret are disabled.

    Restart coturn after changing the configuration file.

    Remember to restore the original settings to go back to testing with Matrix clients!

  If the TURN server is working correctly, you should see at least one relay entry in the results.

Delegation of incoming federation traffic

In the following documentation, we use the term server_name to refer to that setting in your homeserver configuration file. It appears at the ends of user ids, and tells other homeservers where they can find your server.

By default, other homeservers will expect to be able to reach yours via your server_name, on port 8448. For example, if you set your server_name to example.com (so that your user names look like @user:example.com), other servers will try to connect to yours at https://example.com:8448/.

Delegation is a Matrix feature allowing a homeserver admin to retain a server_name of example.com so that user IDs, room aliases, etc continue to look like *:example.com, whilst having federation traffic routed to a different server and/or port (e.g. synapse.example.com:443).

.well-known delegation

To use this method, you need to be able to configure the server at https://<server_name> to serve a file at https://<server_name>/.well-known/matrix/server. There are two ways to do this, shown below.

Note that the .well-known file is hosted on the default port for https (port 443).

External server

For maximum flexibility, you need to configure an external server such as nginx, Apache or HAProxy to serve the https://<server_name>/.well-known/matrix/server file. Setting up such a server is out of the scope of this documentation, but note that it is often possible to configure your reverse proxy for this.

The URL https://<server_name>/.well-known/matrix/server should be configured return a JSON structure containing the key m.server like this:

{
    "m.server": "<synapse.server.name>[:<yourport>]"
}

In our example (where we want federation traffic to be routed to https://synapse.example.com, on port 443), this would mean that https://example.com/.well-known/matrix/server should return:

{
    "m.server": "synapse.example.com:443"
}

Note, specifying a port is optional. If no port is specified, then it defaults to 8448.

Serving a .well-known/matrix/server file with Synapse

If you are able to set up your domain so that https://<server_name> is routed to Synapse (i.e., the only change needed is to direct federation traffic to port 443 instead of port 8448), then it is possible to configure Synapse to serve a suitable .well-known/matrix/server file. To do so, add the following to your homeserver.yaml file:

serve_server_wellknown: true

Note: this only works if https://<server_name> is routed to Synapse, so is generally not suitable if Synapse is hosted at a subdomain such as https://synapse.example.com.

SRV DNS record delegation

It is also possible to do delegation using a SRV DNS record. However, that is generally not recommended, as it can be difficult to configure the TLS certificates correctly in this case, and it offers little advantage over .well-known delegation.

However, if you really need it, you can find some documentation on what such a record should look like and how Synapse will use it in the Matrix specification.

Delegation FAQ

When do I need delegation?

If your homeserver's APIs are accessible on the default federation port (8448) and the domain your server_name points to, you do not need any delegation.

For instance, if you registered example.com and pointed its DNS A record at a fresh server, you could install Synapse on that host, giving it a server_name of example.com, and once a reverse proxy has been set up to proxy all requests sent to the port 8448 and serve TLS certificates for example.com, you wouldn't need any delegation set up.

However, if your homeserver's APIs aren't accessible on port 8448 and on the domain server_name points to, you will need to let other servers know how to find it using delegation.

Should I use a reverse proxy for federation traffic?

Generally, using a reverse proxy for both the federation and client traffic is a good idea, since it saves handling TLS traffic in Synapse. See the reverse proxy documentation for information on setting up a reverse proxy.

Upgrading Synapse

Before upgrading check if any special steps are required to upgrade from the version you currently have installed to the current version of Synapse. The extra instructions that may be required are listed later in this document.

• Check that your versions of Python and PostgreSQL are still supported.

  Synapse follows upstream lifecycles for Python and PostgreSQL, and removes support for versions which are no longer maintained.

  The website https://endoflife.date also offers convenient summaries.

• If Synapse was installed using prebuilt packages, you will need to follow the normal process for upgrading those packages.

• If Synapse was installed using pip then upgrade to the latest version by running:

  pip install --upgrade matrix-synapse
  

• If Synapse was installed from source, then:

  1. Obtain the latest version of the source code. Git users can run git pull to do this.

  2. If you're running Synapse in a virtualenv, make sure to activate it before upgrading. For example, if Synapse is installed in a virtualenv in ~/synapse/env then run:

     source ~/synapse/env/bin/activate
     pip install --upgrade .
     

     Include any relevant extras between square brackets, e.g. pip install --upgrade ".[postgres,oidc]".

  3. If you're using poetry to manage a Synapse installation, run:

     poetry install
     

     Include any relevant extras with --extras, e.g. poetry install --extras postgres --extras oidc. It's probably easiest to run poetry install --extras all.

  4. Restart Synapse:

     synctl restart
     

To check whether your update was successful, you can check the running server version with:

# you may need to replace 'localhost:8008' if synapse is not configured
# to listen on port 8008.

curl http://localhost:8008/_synapse/admin/v1/server_version

Rolling back to older versions

Rolling back to previous releases can be difficult, due to database schema changes between releases. Where we have been able to test the rollback process, this will be noted below.

In general, you will need to undo any changes made during the upgrade process, for example:

• pip:

  source env/bin/activate
  # replace `1.3.0` accordingly:
  pip install matrix-synapse==1.3.0
  

• Debian:

  # replace `1.3.0` and `stretch` accordingly:
  wget https://packages.matrix.org/debian/pool/main/m/matrix-synapse-py3/matrix-synapse-py3_1.3.0+stretch1_amd64.deb
  dpkg -i matrix-synapse-py3_1.3.0+stretch1_amd64.deb
  

Upgrading to v1.67.0

Direct TCP replication is no longer supported: migrate to Redis

Redis support was added in v1.13.0 with it becoming the recommended method in v1.18.0. It replaced the old direct TCP connections (which was deprecated as of v1.18.0) to the main process. With Redis, rather than all the workers connecting to the main process, all the workers and the main process connect to Redis, which relays replication commands between processes. This can give a significant CPU saving on the main process and is a prerequisite for upcoming performance improvements.

To migrate to Redis add the redis config, and remove the TCP replication listener from config of the master and worker_replication_port from worker config. Note that a HTTP listener with a replication resource is still required.

Minimum version of Poetry is now v1.2.0

The minimum supported version of poetry is now 1.2. This should only affect those installing from a source checkout.

Upgrading to v1.66.0

Delegation of email validation no longer supported

As of this version, Synapse no longer allows the tasks of verifying email address ownership, and password reset confirmation, to be delegated to an identity server. This removal was previously planned for Synapse 1.64.0, but was delayed until now to give homeserver administrators more notice of the change.

To continue to allow users to add email addresses to their homeserver accounts, and perform password resets, make sure that Synapse is configured with a working email server in the email configuration section (including, at a minimum, a notif_from setting.)

Specifying an email setting under account_threepid_delegates will now cause an error at startup.

Upgrading to v1.64.0

Deprecation of the ability to delegate e-mail verification to identity servers

Synapse v1.66.0 will remove the ability to delegate the tasks of verifying email address ownership, and password reset confirmation, to an identity server.

If you require your homeserver to verify e-mail addresses or to support password resets via e-mail, please configure your homeserver with SMTP access so that it can send e-mails on its own behalf. Consult the configuration documentation for more information.

The option that will be removed is account_threepid_delegates.email.

Changes to the event replication streams

Synapse now includes a flag indicating if an event is an outlier when replicating it to other workers. This is a forwards- and backwards-incompatible change: v1.63 and workers cannot process events replicated by v1.64 workers, and vice versa.

Once all workers are upgraded to v1.64 (or downgraded to v1.63), event replication will resume as normal.

frozendict release

frozendict 2.3.3 has recently been released, which fixes a memory leak that occurs during /sync requests. We advise server administrators who installed Synapse via pip to upgrade frozendict with pip install --upgrade frozendict. The Docker image matrixdotorg/synapse and the Debian packages from packages.matrix.org already include the updated library.

Upgrading to v1.62.0

New signatures for spam checker callbacks

As a followup to changes in v1.60.0, the following spam-checker callbacks have changed signature:

• user_may_join_room
• user_may_invite
• user_may_send_3pid_invite
• user_may_create_room
• user_may_create_room_alias
• user_may_publish_room
• check_media_file_for_spam

For each of these methods, the previous callback signature has been deprecated.

Whereas callbacks used to return bool, they should now return Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes"].

For instance, if your module implements user_may_join_room as follows:

async def user_may_join_room(self, user_id: str, room_id: str, is_invited: bool)
    if ...:
        # Request is spam
        return False
    # Request is not spam
    return True

you should rewrite it as follows:

async def user_may_join_room(self, user_id: str, room_id: str, is_invited: bool)
    if ...:
        # Request is spam, mark it as forbidden (you may use some more precise error
        # code if it is useful).
        return synapse.module_api.errors.Codes.FORBIDDEN
    # Request is not spam, mark it as such.
    return synapse.module_api.NOT_SPAM

Upgrading to v1.61.0

Removal of deprecated community/groups

This release of Synapse will remove deprecated community/groups from codebase.

Worker endpoints

For those who have deployed workers, following worker endpoints will no longer exist and they can be removed from the reverse proxy configuration:

• ^/_matrix/federation/v1/get_groups_publicised$
• ^/_matrix/client/(r0|v3|unstable)/joined_groups$
• ^/_matrix/client/(r0|v3|unstable)/publicised_groups$
• ^/_matrix/client/(r0|v3|unstable)/publicised_groups/
• ^/_matrix/federation/v1/groups/
• ^/_matrix/client/(r0|v3|unstable)/groups/

Upgrading to v1.60.0

Adding a new unique index to state_group_edges could fail if your database is corrupted

This release of Synapse will add a unique index to the state_group_edges table, in order to prevent accidentally introducing duplicate information (for example, because a database backup was restored multiple times).

Duplicate rows being present in this table could cause drastic performance problems; see issue 11779 for more details.

If your Synapse database already has had duplicate rows introduced into this table, this could fail, with either of these errors:

On Postgres:

synapse.storage.background_updates - 623 - INFO - background_updates-0 - Adding index state_group_edges_unique_idx to state_group_edges
synapse.storage.background_updates - 282 - ERROR - background_updates-0 - Error doing update
...
psycopg2.errors.UniqueViolation: could not create unique index "state_group_edges_unique_idx"
DETAIL:  Key (state_group, prev_state_group)=(2, 1) is duplicated.

(The numbers may be different.)

On SQLite:

synapse.storage.background_updates - 623 - INFO - background_updates-0 - Adding index state_group_edges_unique_idx to state_group_edges
synapse.storage.background_updates - 282 - ERROR - background_updates-0 - Error doing update
...
sqlite3.IntegrityError: UNIQUE constraint failed: state_group_edges.state_group, state_group_edges.prev_state_group

BEGIN;
DELETE FROM state_group_edges WHERE (ctid, state_group, prev_state_group) IN (
  SELECT row_id, state_group, prev_state_group
  FROM (
    SELECT
      ctid AS row_id,
      MIN(ctid) OVER (PARTITION BY state_group, prev_state_group) AS min_row_id,
      state_group,
      prev_state_group
    FROM state_group_edges
  ) AS t1
  WHERE row_id <> min_row_id
);
COMMIT;

BEGIN;
DELETE FROM state_group_edges WHERE (rowid, state_group, prev_state_group) IN (
  SELECT row_id, state_group, prev_state_group
  FROM (
    SELECT
      rowid AS row_id,
      MIN(rowid) OVER (PARTITION BY state_group, prev_state_group) AS min_row_id,
      state_group,
      prev_state_group
    FROM state_group_edges
  )
  WHERE row_id <> min_row_id
);
COMMIT;

New signature for the spam checker callback check_event_for_spam

The previous signature has been deprecated.

Whereas check_event_for_spam callbacks used to return Union[str, bool], they should now return Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes"].

This is part of an ongoing refactoring of the SpamChecker API to make it less ambiguous and more powerful.

If your module implements check_event_for_spam as follows:

async def check_event_for_spam(event):
    if ...:
        # Event is spam
        return True
    # Event is not spam
    return False

you should rewrite it as follows:

async def check_event_for_spam(event):
    if ...:
        # Event is spam, mark it as forbidden (you may use some more precise error
        # code if it is useful).
        return synapse.module_api.errors.Codes.FORBIDDEN
    # Event is not spam, mark it as such.
    return synapse.module_api.NOT_SPAM

Upgrading to v1.59.0

Device name lookup over federation has been disabled by default

The names of user devices are no longer visible to users on other homeservers by default. Device IDs are unaffected, as these are necessary to facilitate end-to-end encryption.

To re-enable this functionality, set the allow_device_name_lookup_over_federation homeserver config option to true.

Deprecation of the synapse.app.appservice and synapse.app.user_dir worker application types

The synapse.app.appservice worker application type allowed you to configure a single worker to use to notify application services of new events, as long as this functionality was disabled on the main process with notify_appservices: False. Further, the synapse.app.user_dir worker application type allowed you to configure a single worker to be responsible for updating the user directory, as long as this was disabled on the main process with update_user_directory: False.

To unify Synapse's worker types, the synapse.app.appservice worker application type and the notify_appservices configuration option have been deprecated. The synapse.app.user_dir worker application type and update_user_directory configuration option have also been deprecated.

To get the same functionality as was provided by the deprecated options, it's now recommended that the synapse.app.generic_worker worker application type is used and that the notify_appservices_from_worker and/or update_user_directory_from_worker options are set to the name of a worker.

For the time being, the old options can be used alongside the new options to make it easier to transition between the two configurations, however please note that:

• the options must not contradict each other (otherwise Synapse won't start); and
• the notify_appservices and update_user_directory options will be removed in a future release of Synapse.

Please see the Notifying Application Services and Updating the User Directory sections of the worker documentation for more information.

Upgrading to v1.58.0

Groups/communities feature has been disabled by default

The non-standard groups/communities feature in Synapse has been disabled by default and will be removed in Synapse v1.61.0.

Upgrading to v1.57.0

Changes to database schema for application services

Synapse v1.57.0 includes a change to the way transaction IDs are managed for application services. If your deployment uses a dedicated worker for application service traffic, it must be stopped when the database is upgraded (which normally happens when the main process is upgraded), to ensure the change is made safely without any risk of reusing transaction IDs.

Deployments which do not use separate worker processes can be upgraded as normal. Similarly, deployments where no application services are in use can be upgraded as normal.

**********************************************************************************
 Error during initialisation:

 Postgres sequence 'application_services_txn_id_seq' is inconsistent with associated
 table 'application_services_txns'. This can happen if Synapse has been downgraded and
 then upgraded again, or due to a bad migration.

 To fix this error, shut down Synapse (including any and all workers)
 and run the following SQL:

     SELECT setval('application_services_txn_id_seq', (
         SELECT GREATEST(MAX(txn_id), 0) FROM application_services_txns
     ));

 See docs/postgres.md for more information.

 There may be more information in the logs.
**********************************************************************************

Upgrading to v1.56.0

Open registration without verification is now disabled by default

Synapse will refuse to start if registration is enabled without email, captcha, or token-based verification unless the new config flag enable_registration_without_verification is set to "true".

Groups/communities feature has been deprecated

The non-standard groups/communities feature in Synapse has been deprecated and will be disabled by default in Synapse v1.58.0.

You can test disabling it by adding the following to your homeserver configuration:

experimental_features:
  groups_enabled: false

Change in behaviour for PostgreSQL databases with unsafe locale

Synapse now refuses to start when using PostgreSQL with non-C values for COLLATE and CTYPE unless the config flag allow_unsafe_locale, found in the database section of the configuration file, is set to true. See the PostgreSQL documentation for more information and instructions on how to fix a database with incorrect values.

Upgrading to v1.55.0

synctl script has been moved

The synctl script has been made an entry point and no longer exists at the root of Synapse's source tree. If you wish to use synctl to manage your homeserver, you should invoke synctl directly, e.g. synctl start instead of ./synctl start or /path/to/synctl start.

You will need to ensure synctl is on your PATH.

• This is automatically the case when using Debian packages or docker images provided by Matrix.org.
• When installing from a wheel, sdist, or PyPI, a synctl executable is added to your Python installation's bin. This should be on your PATH automatically, though you might need to activate a virtual environment depending on how you installed Synapse.

Compatibility dropped for Mjolnir 1.3.1 and earlier

Synapse v1.55.0 drops support for Mjolnir 1.3.1 and earlier. If you use the Mjolnir module to moderate your homeserver, please upgrade Mjolnir to version 1.3.2 or later before upgrading Synapse.

Upgrading to v1.54.0

Legacy structured logging configuration removal

This release removes support for the structured: true logging configuration which was deprecated in Synapse v1.23.0. If your logging configuration contains structured: true then it should be modified based on the structured logging documentation.

Upgrading to v1.53.0

Dropping support for webclient listeners and non-HTTP(S) web_client_location

Per the deprecation notice in Synapse v1.51.0, listeners of type webclient are no longer supported and configuring them is a now a configuration error.

Configuring a non-HTTP(S) web_client_location configuration is is now a configuration error. Since the webclient listener is no longer supported, this setting only applies to the root path / of Synapse's web server and no longer the /_matrix/client/ path.

Stablisation of MSC3231

The unstable validity-check endpoint for the Registration Tokens feature has been stabilised and moved from:

/_matrix/client/unstable/org.matrix.msc3231/register/org.matrix.msc3231.login.registration_token/validity

to:

/_matrix/client/v1/register/m.login.registration_token/validity

Please update any relevant reverse proxy or firewall configurations appropriately.

Time-based cache expiry is now enabled by default

Formerly, entries in the cache were not evicted regardless of whether they were accessed after storing. This behavior has now changed. By default entries in the cache are now evicted after 30m of not being accessed. To change the default behavior, go to the caches section of the config and change the expire_caches and cache_entry_ttl flags as necessary. Please note that these flags replace the expiry_time flag in the config. The expiry_time flag will still continue to work, but it has been deprecated and will be removed in the future.

Deprecation of capability org.matrix.msc3283.*

The capabilities of MSC3283 from the REST API /_matrix/client/r0/capabilities becomes stable.

The old capabilities

• org.matrix.msc3283.set_displayname,
• org.matrix.msc3283.set_avatar_url and
• org.matrix.msc3283.3pid_changes

are deprecated and scheduled to be removed in Synapse v1.54.0.

The new capabilities

• m.set_displayname,
• m.set_avatar_url and
• m.3pid_changes

are now active by default.

Removal of user_may_create_room_with_invites

As announced with the release of Synapse 1.47.0, the deprecated user_may_create_room_with_invites module callback has been removed.

Modules relying on it can instead implement user_may_invite and use the get_room_state module API to infer whether the invite is happening while creating a room (see this function as an example). Alternately, modules can also implement on_create_room.

Upgrading to v1.52.0

Twisted security release

Note that Twisted 22.1.0 has recently been released, which fixes a security issue within the Twisted library. We do not believe Synapse is affected by this vulnerability, though we advise server administrators who installed Synapse via pip to upgrade Twisted with pip install --upgrade Twisted treq as a matter of good practice. The Docker image matrixdotorg/synapse and the Debian packages from packages.matrix.org are using the updated library.

Upgrading to v1.51.0

Deprecation of webclient listeners and non-HTTP(S) web_client_location

Listeners of type webclient are deprecated and scheduled to be removed in Synapse v1.53.0.

Similarly, a non-HTTP(S) web_client_location configuration is deprecated and will become a configuration error in Synapse v1.53.0.

Upgrading to v1.50.0

Dropping support for old Python and Postgres versions

In line with our deprecation policy, we've dropped support for Python 3.6 and PostgreSQL 9.6, as they are no longer supported upstream.

This release of Synapse requires Python 3.7+ and PostgreSQL 10+.

Upgrading to v1.47.0

Removal of old Room Admin API

The following admin APIs were deprecated in Synapse 1.34 (released on 2021-05-17) and have now been removed:

• POST /_synapse/admin/v1/<room_id>/delete

Any scripts still using the above APIs should be converted to use the Delete Room API.

Deprecation of the user_may_create_room_with_invites module callback

The user_may_create_room_with_invites is deprecated and will be removed in a future version of Synapse. Modules implementing this callback can instead implement user_may_invite and use the get_room_state module API method to infer whether the invite is happening in the context of creating a room.

We plan to remove this callback in January 2022.

Upgrading to v1.45.0

Changes required to media storage provider modules when reading from the Synapse configuration object

Media storage provider modules that read from the Synapse configuration object (i.e. that read the value of hs.config.[...]) now need to specify the configuration section they're reading from. This means that if a module reads the value of e.g. hs.config.media_store_path, it needs to replace it with hs.config.media.media_store_path.

Upgrading to v1.44.0

The URL preview cache is no longer mirrored to storage providers

The url_cache/ and url_cache_thumbnails/ directories in the media store are no longer mirrored to storage providers. These two directories can be safely deleted from any configured storage providers to reclaim space.

Upgrading to v1.43.0

The spaces summary APIs can now be handled by workers

The available worker applications documentation has been updated to reflect that calls to the /spaces, /hierarchy, and /summary endpoints can now be routed to workers for both client API and federation requests.

Upgrading to v1.42.0

Removal of old Room Admin API

The following admin APIs were deprecated in Synapse 1.25 (released on 2021-01-13) and have now been removed:

• POST /_synapse/admin/v1/purge_room
• POST /_synapse/admin/v1/shutdown_room/<room_id>

Any scripts still using the above APIs should be converted to use the Delete Room API.

User-interactive authentication fallback templates can now display errors

This may affect you if you make use of custom HTML templates for the reCAPTCHA or terms fallback pages.

The template is now provided an error variable if the authentication process failed. See the default templates linked above for an example.

Removal of out-of-date email pushers

Users will stop receiving message updates via email for addresses that were once, but not still, linked to their account.

Upgrading to v1.41.0

Add support for routing outbound HTTP requests via a proxy for federation

Since Synapse 1.6.0 (2019-11-26) you can set a proxy for outbound HTTP requests via http_proxy/https_proxy environment variables. This proxy was set for:

• push
• url previews
• phone-home stats
• recaptcha validation
• CAS auth validation
• OpenID Connect
• Federation (checking public key revocation)

In this version we have added support for outbound requests for:

• Outbound federation
• Downloading remote media
• Fetching public keys of other servers

These requests use the same proxy configuration. If you have a proxy configuration we recommend to verify the configuration. It may be necessary to adjust the no_proxy environment variable.

See using a forward proxy with Synapse documentation for details.

Deprecation of template_dir

The template_dir settings in the sso, account_validity and email sections of the configuration file are now deprecated. Server admins should use the new templates.custom_template_directory setting in the configuration file and use one single custom template directory for all aforementioned features. Template file names remain unchanged. See the related documentation for more information and examples.

We plan to remove support for these settings in October 2021.

/_synapse/admin/v1/users/{userId}/media must be handled by media workers

The media repository worker documentation has been updated to reflect that calls to /_synapse/admin/v1/users/{userId}/media must now be handled by media repository workers. This is due to the new DELETE method of this endpoint modifying the media store.

Upgrading to v1.39.0

Deprecation of the current third-party rules module interface

The current third-party rules module interface is deprecated in favour of the new generic modules system introduced in Synapse v1.37.0. Authors of third-party rules modules can refer to this documentation to update their modules. Synapse administrators can refer to this documentation to update their configuration once the modules they are using have been updated.

We plan to remove support for the current third-party rules interface in September 2021.

Upgrading to v1.38.0

Re-indexing of events table on Postgres databases

This release includes a database schema update which requires re-indexing one of the larger tables in the database, events. This could result in increased disk I/O for several hours or days after upgrading while the migration completes. Furthermore, because we have to keep the old indexes until the new indexes are ready, it could result in a significant, temporary, increase in disk space.

To get a rough idea of the disk space required, check the current size of one of the indexes. For example, from a psql shell, run the following sql:

SELECT pg_size_pretty(pg_relation_size('events_order_room'));

We need to rebuild four indexes, so you will need to multiply this result by four to give an estimate of the disk space required. For example, on one particular server:

synapse=# select pg_size_pretty(pg_relation_size('events_order_room'));
 pg_size_pretty
----------------
 288 MB
(1 row)

On this server, it would be wise to ensure that at least 1152MB are free.

The additional disk space will be freed once the migration completes.

SQLite databases are unaffected by this change.

Upgrading to v1.37.0

Deprecation of the current spam checker interface

The current spam checker interface is deprecated in favour of a new generic modules system. Authors of spam checker modules can refer to [this documentation](modules/porting_legacy_module.md to update their modules. Synapse administrators can refer to this documentation to update their configuration once the modules they are using have been updated.

We plan to remove support for the current spam checker interface in August 2021.

More module interfaces will be ported over to this new generic system in future versions of Synapse.

Upgrading to v1.34.0

room_invite_state_types configuration setting

The room_invite_state_types configuration setting has been deprecated and replaced with room_prejoin_state. See the sample configuration file.

If you have set room_invite_state_types to the default value you should simply remove it from your configuration file. The default value used to be:

room_invite_state_types:
   - "m.room.join_rules"
   - "m.room.canonical_alias"
   - "m.room.avatar"
   - "m.room.encryption"
   - "m.room.name"

If you have customised this value, you should remove room_invite_state_types and configure room_prejoin_state instead.

Upgrading to v1.33.0

Account Validity HTML templates can now display a user's expiration date

This may affect you if you have enabled the account validity feature, and have made use of a custom HTML template specified by the account_validity.template_dir or account_validity.account_renewed_html_path Synapse config options.

The template can now accept an expiration_ts variable, which represents the unix timestamp in milliseconds for the future date of which their account has been renewed until. See the default template for an example of usage.

ALso note that a new HTML template, account_previously_renewed.html, has been added. This is is shown to users when they attempt to renew their account with a valid renewal token that has already been used before. The default template contents can been found here, and can also accept an expiration_ts variable. This template replaces the error message users would previously see upon attempting to use a valid renewal token more than once.

Upgrading to v1.32.0

Regression causing connected Prometheus instances to become overwhelmed

This release introduces a regression that can overwhelm connected Prometheus instances. This issue is not present in Synapse v1.32.0rc1.

If you have been affected, please downgrade to 1.31.0. You then may need to remove excess writeahead logs in order for Prometheus to recover. Instructions for doing so are provided here.

Dropping support for old Python, Postgres and SQLite versions

In line with our deprecation policy, we've dropped support for Python 3.5 and PostgreSQL 9.5, as they are no longer supported upstream.

This release of Synapse requires Python 3.6+ and PostgresSQL 9.6+ or SQLite 3.22+.

Removal of old List Accounts Admin API

The deprecated v1 "list accounts" admin API (GET /_synapse/admin/v1/users/<user_id>) has been removed in this version.

The v2 list accounts API has been available since Synapse 1.7.0 (2019-12-13), and is accessible under GET /_synapse/admin/v2/users.

The deprecation of the old endpoint was announced with Synapse 1.28.0 (released on 2021-02-25).

Application Services must use type m.login.application_service when registering users

In compliance with the Application Service spec, Application Services are now required to use the m.login.application_service type when registering users via the /_matrix/client/r0/register endpoint. This behaviour was deprecated in Synapse v1.30.0.

Please ensure your Application Services are up to date.

Upgrading to v1.29.0

Requirement for X-Forwarded-Proto header

When using Synapse with a reverse proxy (in particular, when using the x_forwarded option on an HTTP listener), Synapse now expects to receive an X-Forwarded-Proto header on incoming HTTP requests. If it is not set, Synapse will log a warning on each received request.

To avoid the warning, administrators using a reverse proxy should ensure that the reverse proxy sets X-Forwarded-Proto header to https or http to indicate the protocol used by the client.

Synapse also requires the Host header to be preserved.

See the reverse proxy documentation, where the example configurations have been updated to show how to set these headers.

(Users of Caddy are unaffected, since we believe it sets X-Forwarded-Proto by default.)

Upgrading to v1.27.0

Changes to callback URI for OAuth2 / OpenID Connect and SAML2

This version changes the URI used for callbacks from OAuth2 and SAML2 identity providers:

• If your server is configured for single sign-on via an OpenID Connect or OAuth2 identity provider, you will need to add [synapse public baseurl]/_synapse/client/oidc/callback to the list of permitted "redirect URIs" at the identity provider.

  See the OpenID docs for more information on setting up OpenID Connect.

• If your server is configured for single sign-on via a SAML2 identity provider, you will need to add [synapse public baseurl]/_synapse/client/saml2/authn_response as a permitted "ACS location" (also known as "allowed callback URLs") at the identity provider.

  The "Issuer" in the "AuthnRequest" to the SAML2 identity provider is also updated to [synapse public baseurl]/_synapse/client/saml2/metadata.xml. If your SAML2 identity provider uses this property to validate or otherwise identify Synapse, its configuration will need to be updated to use the new URL. Alternatively you could create a new, separate "EntityDescriptor" in your SAML2 identity provider with the new URLs and leave the URLs in the existing "EntityDescriptor" as they were.

Changes to HTML templates

The HTML templates for SSO and email notifications now have Jinja2's autoescape enabled for files ending in .html, .htm, and .xml. If you have customised these templates and see issues when viewing them you might need to update them. It is expected that most configurations will need no changes.

If you have customised the templates names for these templates, it is recommended to verify they end in .html to ensure autoescape is enabled.

The above applies to the following templates:

• add_threepid.html
• add_threepid_failure.html
• add_threepid_success.html
• notice_expiry.html
• notice_expiry.html
• notif_mail.html (which, by default, includes room.html and notif.html)
• password_reset.html
• password_reset_confirmation.html
• password_reset_failure.html
• password_reset_success.html
• registration.html
• registration_failure.html
• registration_success.html
• sso_account_deactivated.html
• sso_auth_bad_user.html
• sso_auth_confirm.html
• sso_auth_success.html
• sso_error.html
• sso_login_idp_picker.html
• sso_redirect_confirm.html

Upgrading to v1.26.0

Rolling back to v1.25.0 after a failed upgrade

v1.26.0 includes a lot of large changes. If something problematic occurs, you may want to roll-back to a previous version of Synapse. Because v1.26.0 also includes a new database schema version, reverting that version is also required alongside the generic rollback instructions mentioned above. In short, to roll back to v1.25.0 you need to:

1. Stop the server

2. Decrease the schema version in the database:

   UPDATE schema_version SET version = 58;
   

3. Delete the ignored users & chain cover data:

   DROP TABLE IF EXISTS ignored_users;
   UPDATE rooms SET has_auth_chain_index = false;
   

   For PostgreSQL run:

   TRUNCATE event_auth_chain_links;
   TRUNCATE event_auth_chains;
   

   For SQLite run:

   DELETE FROM event_auth_chain_links;
   DELETE FROM event_auth_chains;
   

4. Mark the deltas as not run (so they will re-run on upgrade).

   DELETE FROM applied_schema_deltas WHERE version = 59 AND file = "59/01ignored_user.py";
   DELETE FROM applied_schema_deltas WHERE version = 59 AND file = "59/06chain_cover_index.sql";
   

5. Downgrade Synapse by following the instructions for your installation method in the "Rolling back to older versions" section above.

Upgrading to v1.25.0

Last release supporting Python 3.5

This is the last release of Synapse which guarantees support with Python 3.5, which passed its upstream End of Life date several months ago.

We will attempt to maintain support through March 2021, but without guarantees.

In the future, Synapse will follow upstream schedules for ending support of older versions of Python and PostgreSQL. Please upgrade to at least Python 3.6 and PostgreSQL 9.6 as soon as possible.

Blacklisting IP ranges

Synapse v1.25.0 includes new settings, ip_range_blacklist and ip_range_whitelist, for controlling outgoing requests from Synapse for federation, identity servers, push, and for checking key validity for third-party invite events. The previous setting, federation_ip_range_blacklist, is deprecated. The new ip_range_blacklist defaults to private IP ranges if it is not defined.

If you have never customised federation_ip_range_blacklist it is recommended that you remove that setting.

If you have customised federation_ip_range_blacklist you should update the setting name to ip_range_blacklist.

If you have a custom push server that is reached via private IP space you may need to customise ip_range_blacklist or ip_range_whitelist.

Upgrading to v1.24.0

Custom OpenID Connect mapping provider breaking change

This release allows the OpenID Connect mapping provider to perform normalisation of the localpart of the Matrix ID. This allows for the mapping provider to specify different algorithms, instead of the default way.

If your Synapse configuration uses a custom mapping provider (oidc_config.user_mapping_provider.module is specified and not equal to synapse.handlers.oidc_handler.JinjaOidcMappingProvider) then you must ensure that map_user_attributes of the mapping provider performs some normalisation of the localpart returned. To match previous behaviour you can use the map_username_to_mxid_localpart function provided by Synapse. An example is shown below:

from synapse.types import map_username_to_mxid_localpart

class MyMappingProvider:
    def map_user_attributes(self, userinfo, token):
        # ... your custom logic ...
        sso_user_id = ...
        localpart = map_username_to_mxid_localpart(sso_user_id)

        return {"localpart": localpart}

Removal historical Synapse Admin API

Historically, the Synapse Admin API has been accessible under:

• /_matrix/client/api/v1/admin
• /_matrix/client/unstable/admin
• /_matrix/client/r0/admin
• /_synapse/admin/v1

The endpoints with /_matrix/client/* prefixes have been removed as of v1.24.0. The Admin API is now only accessible under:

• /_synapse/admin/v1

The only exception is the /admin/whois endpoint, which is also available via the client-server API.

The deprecation of the old endpoints was announced with Synapse 1.20.0 (released on 2020-09-22) and makes it easier for homeserver admins to lock down external access to the Admin API endpoints.

Upgrading to v1.23.0

Structured logging configuration breaking changes

This release deprecates use of the structured: true logging configuration for structured logging. If your logging configuration contains structured: true then it should be modified based on the structured logging documentation.

The structured and drains logging options are now deprecated and should be replaced by standard logging configuration of handlers and formatters.

A future will release of Synapse will make using structured: true an error.

Upgrading to v1.22.0

ThirdPartyEventRules breaking changes

This release introduces a backwards-incompatible change to modules making use of ThirdPartyEventRules in Synapse. If you make use of a module defined under the third_party_event_rules config option, please make sure it is updated to handle the below change:

The http_client argument is no longer passed to modules as they are initialised. Instead, modules are expected to make use of the http_client property on the ModuleApi class. Modules are now passed a module_api argument during initialisation, which is an instance of ModuleApi. ModuleApi instances have a http_client property which acts the same as the http_client argument previously passed to ThirdPartyEventRules modules.

Upgrading to v1.21.0

Forwarding /_synapse/client through your reverse proxy

The reverse proxy documentation has been updated to include reverse proxy directives for /_synapse/client/* endpoints. As the user password reset flow now uses endpoints under this prefix, you must update your reverse proxy configurations for user password reset to work.

Additionally, note that the Synapse worker documentation has been updated to

: state that the /_synapse/client/password_reset/email/submit_token endpoint can be handled

by all workers. If you make use of Synapse's worker feature, please update your reverse proxy configuration to reflect this change.

New HTML templates

A new HTML template, password_reset_confirmation.html, has been added to the synapse/res/templates directory. If you are using a custom template directory, you may want to copy the template over and modify it.

Note that as of v1.20.0, templates do not need to be included in custom template directories for Synapse to start. The default templates will be used if a custom template cannot be found.

This page will appear to the user after clicking a password reset link that has been emailed to them.

To complete password reset, the page must include a way to make a POST request to /_synapse/client/password_reset/{medium}/submit_token with the query parameters from the original link, presented as a URL-encoded form. See the file itself for more details.

Updated Single Sign-on HTML Templates

The saml_error.html template was removed from Synapse and replaced with the sso_error.html template. If your Synapse is configured to use SAML and a custom sso_redirect_confirm_template_dir configuration then any customisations of the saml_error.html template will need to be merged into the sso_error.html template. These templates are similar, but the parameters are slightly different:

• The msg parameter should be renamed to error_description.
• There is no longer a code parameter for the response code.
• A string error parameter is available that includes a short hint of why a user is seeing the error page.

Upgrading to v1.18.0

Docker -py3 suffix will be removed in future versions

From 10th August 2020, we will no longer publish Docker images with the -py3 tag suffix. The images tagged with the -py3 suffix have been identical to the non-suffixed tags since release 0.99.0, and the suffix is obsolete.

On 10th August, we will remove the latest-py3 tag. Existing per-release tags (such as v1.18.0-py3 will not be removed, but no new -py3 tags will be added.

Scripts relying on the -py3 suffix will need to be updated.

Redis replication is now recommended in lieu of TCP replication

When setting up worker processes, we now recommend the use of a Redis server for replication. The old direct TCP connection method is deprecated and will be removed in a future release. See the worker documentation for more details.

Upgrading to v1.14.0

This version includes a database update which is run as part of the upgrade, and which may take a couple of minutes in the case of a large server. Synapse will not respond to HTTP requests while this update is taking place.

Upgrading to v1.13.0

Incorrect database migration in old synapse versions

A bug was introduced in Synapse 1.4.0 which could cause the room directory to be incomplete or empty if Synapse was upgraded directly from v1.2.1 or earlier, to versions between v1.4.0 and v1.12.x.

This will not be a problem for Synapse installations which were:

: - created at v1.4.0 or later, - upgraded via v1.3.x, or - upgraded straight from v1.2.1 or earlier to v1.13.0 or later.

If completeness of the room directory is a concern, installations which are affected can be repaired as follows:

1. Run the following sql from a psql or sqlite3 console:

   INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES
      ('populate_stats_process_rooms', '{}', 'current_state_events_membership');
   
   INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES
      ('populate_stats_process_users', '{}', 'populate_stats_process_rooms');
   

2. Restart synapse.

New Single Sign-on HTML Templates

New templates (sso_auth_confirm.html, sso_auth_success.html, and sso_account_deactivated.html) were added to Synapse. If your Synapse is configured to use SSO and a custom sso_redirect_confirm_template_dir configuration then these templates will need to be copied from synapse/res/templates into that directory.

Synapse SSO Plugins Method Deprecation

Plugins using the complete_sso_login method of synapse.module_api.ModuleApi should update to using the async/await version complete_sso_login_async which includes additional checks. The non-async version is considered deprecated.

Rolling back to v1.12.4 after a failed upgrade

v1.13.0 includes a lot of large changes. If something problematic occurs, you may want to roll-back to a previous version of Synapse. Because v1.13.0 also includes a new database schema version, reverting that version is also required alongside the generic rollback instructions mentioned above. In short, to roll back to v1.12.4 you need to:

1. Stop the server

2. Decrease the schema version in the database:

   UPDATE schema_version SET version = 57;
   

3. Downgrade Synapse by following the instructions for your installation method in the "Rolling back to older versions" section above.

Upgrading to v1.12.0

This version includes a database update which is run as part of the upgrade, and which may take some time (several hours in the case of a large server). Synapse will not respond to HTTP requests while this update is taking place.

This is only likely to be a problem in the case of a server which is participating in many rooms.

0. As with all upgrades, it is recommended that you have a recent backup of your database which can be used for recovery in the event of any problems.

1. As an initial check to see if you will be affected, you can try running the following query from the psql or sqlite3 console. It is safe to run it while Synapse is still running.

   SELECT MAX(q.v) FROM (
     SELECT (
       SELECT ej.json AS v
       FROM state_events se INNER JOIN event_json ej USING (event_id)
       WHERE se.room_id=rooms.room_id AND se.type='m.room.create' AND se.state_key=''
       LIMIT 1
     ) FROM rooms WHERE rooms.room_version IS NULL
   ) q;
   

   This query will take about the same amount of time as the upgrade process: ie, if it takes 5 minutes, then it is likely that Synapse will be unresponsive for 5 minutes during the upgrade.

   If you consider an outage of this duration to be acceptable, no further action is necessary and you can simply start Synapse 1.12.0.

   If you would prefer to reduce the downtime, continue with the steps below.

2. The easiest workaround for this issue is to manually create a new index before upgrading. On PostgreSQL, his can be done as follows:

   CREATE INDEX CONCURRENTLY tmp_upgrade_1_12_0_index
   ON state_events(room_id) WHERE type = 'm.room.create';
   

   The above query may take some time, but is also safe to run while Synapse is running.

   We assume that no SQLite users have databases large enough to be affected. If you are affected, you can run a similar query, omitting the CONCURRENTLY keyword. Note however that this operation may in itself cause Synapse to stop running for some time. Synapse admins are reminded that SQLite is not recommended for use outside a test environment.

3. Once the index has been created, the SELECT query in step 1 above should complete quickly. It is therefore safe to upgrade to Synapse 1.12.0.

4. Once Synapse 1.12.0 has successfully started and is responding to HTTP requests, the temporary index can be removed:

   DROP INDEX tmp_upgrade_1_12_0_index;
   

Upgrading to v1.10.0

Synapse will now log a warning on start up if used with a PostgreSQL database that has a non-recommended locale set.

See Postgres for details.

Upgrading to v1.8.0

Specifying a log_file config option will now cause Synapse to refuse to start, and should be replaced by with the log_config option. Support for the log_file option was removed in v1.3.0 and has since had no effect.

Upgrading to v1.7.0

In an attempt to configure Synapse in a privacy preserving way, the default behaviours of allow_public_rooms_without_auth and allow_public_rooms_over_federation have been inverted. This means that by default, only authenticated users querying the Client/Server API will be able to query the room directory, and relatedly that the server will not share room directory information with other servers over federation.

If your installation does not explicitly set these settings one way or the other and you want either setting to be true then it will necessary to update your homeserver configuration file accordingly.

For more details on the surrounding context see our explainer.

Upgrading to v1.5.0

This release includes a database migration which may take several minutes to complete if there are a large number (more than a million or so) of entries in the devices table. This is only likely to a be a problem on very large installations.

Upgrading to v1.4.0

New custom templates

If you have configured a custom template directory with the email.template_dir option, be aware that there are new templates regarding registration and threepid management (see below) that must be included.

• registration.html and registration.txt
• registration_success.html and registration_failure.html
• add_threepid.html and add_threepid.txt
• add_threepid_failure.html and add_threepid_success.html

Synapse will expect these files to exist inside the configured template directory, and will fail to start if they are absent. To view the default templates, see synapse/res/templates.

3pid verification changes

Note: As of this release, users will be unable to add phone numbers or email addresses to their accounts, without changes to the Synapse configuration. This includes adding an email address during registration.

It is possible for a user to associate an email address or phone number with their account, for a number of reasons:

• for use when logging in, as an alternative to the user id.
• in the case of email, as an alternative contact to help with account recovery.
• in the case of email, to receive notifications of missed messages.

Before an email address or phone number can be added to a user's account, or before such an address is used to carry out a password-reset, Synapse must confirm the operation with the owner of the email address or phone number. It does this by sending an email or text giving the user a link or token to confirm receipt. This process is known as '3pid verification'. ('3pid', or 'threepid', stands for third-party identifier, and we use it to refer to external identifiers such as email addresses and phone numbers.)

Previous versions of Synapse delegated the task of 3pid verification to an identity server by default. In most cases this server is vector.im or matrix.org.

In Synapse 1.4.0, for security and privacy reasons, the homeserver will no longer delegate this task to an identity server by default. Instead, the server administrator will need to explicitly decide how they would like the verification messages to be sent.

In the medium term, the vector.im and matrix.org identity servers will disable support for delegated 3pid verification entirely. However, in order to ease the transition, they will retain the capability for a limited period. Delegated email verification will be disabled on Monday 2nd December 2019 (giving roughly 2 months notice). Disabling delegated SMS verification will follow some time after that once SMS verification support lands in Synapse.

Once delegated 3pid verification support has been disabled in the vector.im and matrix.org identity servers, all Synapse versions that depend on those instances will be unable to verify email and phone numbers through them. There are no imminent plans to remove delegated 3pid verification from Sydent generally. (Sydent is the identity server project that backs the vector.im and matrix.org instances).

Email

Following upgrade, to continue verifying email (e.g. as part of the registration process), admins can either:-

• Configure Synapse to use an email server.
• Run or choose an identity server which allows delegated email verification and delegate to it.

Configure SMTP in Synapse

To configure an SMTP server for Synapse, modify the configuration section headed email, and be sure to have at least the smtp_host, smtp_port and notif_from fields filled out.

You may also need to set smtp_user, smtp_pass, and require_transport_security.

See the sample configuration file for more details on these settings.

Delegate email to an identity server

Some admins will wish to continue using email verification as part of the registration process, but will not immediately have an appropriate SMTP server at hand.

To this end, we will continue to support email verification delegation via the vector.im and matrix.org identity servers for two months. Support for delegated email verification will be disabled on Monday 2nd December.

The account_threepid_delegates dictionary defines whether the homeserver should delegate an external server (typically an identity server) to handle sending confirmation messages via email and SMS.

So to delegate email verification, in homeserver.yaml, set account_threepid_delegates.email to the base URL of an identity server. For example:

account_threepid_delegates:
    email: https://example.com     # Delegate email sending to example.com

Note that account_threepid_delegates.email replaces the deprecated email.trust_identity_server_for_password_resets: if email.trust_identity_server_for_password_resets is set to true, and account_threepid_delegates.email is not set, then the first entry in trusted_third_party_id_servers will be used as the account_threepid_delegate for email. This is to ensure compatibility with existing Synapse installs that set up external server handling for these tasks before v1.4.0. If email.trust_identity_server_for_password_resets is true and no trusted identity server domains are configured, Synapse will report an error and refuse to start.

If email.trust_identity_server_for_password_resets is false or absent and no email delegate is configured in account_threepid_delegates, then Synapse will send email verification messages itself, using the configured SMTP server (see above). that type.

Phone numbers

Synapse does not support phone-number verification itself, so the only way to maintain the ability for users to add phone numbers to their accounts will be by continuing to delegate phone number verification to the matrix.org and vector.im identity servers (or another identity server that supports SMS sending).

The account_threepid_delegates dictionary defines whether the homeserver should delegate an external server (typically an identity server) to handle sending confirmation messages via email and SMS.

So to delegate phone number verification, in homeserver.yaml, set account_threepid_delegates.msisdn to the base URL of an identity server. For example:

account_threepid_delegates:
    msisdn: https://example.com     # Delegate sms sending to example.com

The matrix.org and vector.im identity servers will continue to support delegated phone number verification via SMS until such time as it is possible for admins to configure their servers to perform phone number verification directly. More details will follow in a future release.

Rolling back to v1.3.1

If you encounter problems with v1.4.0, it should be possible to roll back to v1.3.1, subject to the following:

• The 'room statistics' engine was heavily reworked in this release (see #5971), including significant changes to the database schema, which are not easily reverted. This will cause the room statistics engine to stop updating when you downgrade.

  The room statistics are essentially unused in v1.3.1 (in future versions of Synapse, they will be used to populate the room directory), so there should be no loss of functionality. However, the statistics engine will write errors to the logs, which can be avoided by setting the following in homeserver.yaml:

  stats:
    enabled: false
  

  Don't forget to re-enable it when you upgrade again, in preparation for its use in the room directory!

Upgrading to v1.2.0

Some counter metrics have been renamed, with the old names deprecated. See the metrics documentation for details.

Upgrading to v1.1.0

Synapse v1.1.0 removes support for older Python and PostgreSQL versions, as outlined in our deprecation notice.

Minimum Python Version

Synapse v1.1.0 has a minimum Python requirement of Python 3.5. Python 3.6 or Python 3.7 are recommended as they have improved internal string handling, significantly reducing memory usage.

If you use current versions of the Matrix.org-distributed Debian packages or Docker images, action is not required.

If you install Synapse in a Python virtual environment, please see "Upgrading to v0.34.0" for notes on setting up a new virtualenv under Python 3.

Minimum PostgreSQL Version

If using PostgreSQL under Synapse, you will need to use PostgreSQL 9.5 or above. Please see the PostgreSQL documentation for more details on upgrading your database.

Upgrading to v1.0

Validation of TLS certificates

Synapse v1.0 is the first release to enforce validation of TLS certificates for the federation API. It is therefore essential that your certificates are correctly configured.

Note, v1.0 installations will also no longer be able to federate with servers that have not correctly configured their certificates.

In rare cases, it may be desirable to disable certificate checking: for example, it might be essential to be able to federate with a given legacy server in a closed federation. This can be done in one of two ways:-

• Configure the global switch federation_verify_certificates to false.
• Configure a whitelist of server domains to trust via federation_certificate_verification_whitelist.

See the sample configuration file for more details on these settings.

Email

When a user requests a password reset, Synapse will send an email to the user to confirm the request.

Previous versions of Synapse delegated the job of sending this email to an identity server. If the identity server was somehow malicious or became compromised, it would be theoretically possible to hijack an account through this means.

Therefore, by default, Synapse v1.0 will send the confirmation email itself. If Synapse is not configured with an SMTP server, password reset via email will be disabled.

To configure an SMTP server for Synapse, modify the configuration section headed email, and be sure to have at least the smtp_host, smtp_port and notif_from fields filled out. You may also need to set smtp_user, smtp_pass, and require_transport_security.

If you are absolutely certain that you wish to continue using an identity server for password resets, set trust_identity_server_for_password_resets to true.

See the sample configuration file for more details on these settings.

New email templates

Some new templates have been added to the default template directory for the purpose of the homeserver sending its own password reset emails. If you have configured a custom template_dir in your Synapse config, these files will need to be added.

password_reset.html and password_reset.txt are HTML and plain text templates respectively that contain the contents of what will be emailed to the user upon attempting to reset their password via email. password_reset_success.html and password_reset_failure.html are HTML files that the content of which (assuming no redirect URL is set) will be shown to the user after they attempt to click the link in the email sent to them.

Upgrading to v0.99.0

Please be aware that, before Synapse v1.0 is released around March 2019, you will need to replace any self-signed certificates with those verified by a root CA. Information on how to do so can be found at the ACME docs.

Upgrading to v0.34.0

1. This release is the first to fully support Python 3. Synapse will now run on Python versions 3.5, or 3.6 (as well as 2.7). We recommend switching to Python 3, as it has been shown to give performance improvements.

   For users who have installed Synapse into a virtualenv, we recommend doing this by creating a new virtualenv. For example:

   virtualenv -p python3 ~/synapse/env3
   source ~/synapse/env3/bin/activate
   pip install matrix-synapse
   

   You can then start synapse as normal, having activated the new virtualenv:

   cd ~/synapse
   source env3/bin/activate
   synctl start
   

   Users who have installed from distribution packages should see the relevant package documentation. See below for notes on Debian packages.

   • When upgrading to Python 3, you must make sure that your log files are configured as UTF-8, by adding encoding: utf8 to the RotatingFileHandler configuration (if you have one) in your <server>.log.config file. For example, if your log.config file contains:

     handlers:
       file:
         class: logging.handlers.RotatingFileHandler
         formatter: precise
         filename: homeserver.log
         maxBytes: 104857600
         backupCount: 10
         filters: [context]
       console:
         class: logging.StreamHandler
         formatter: precise
         filters: [context]
     

     Then you should update this to be:

     handlers:
       file:
         class: logging.handlers.RotatingFileHandler
         formatter: precise
         filename: homeserver.log
         maxBytes: 104857600
         backupCount: 10
         filters: [context]
         encoding: utf8
       console:
         class: logging.StreamHandler
         formatter: precise
         filters: [context]
     

     There is no need to revert this change if downgrading to Python 2.

   We are also making available Debian packages which will run Synapse on Python 3. You can switch to these packages with apt-get install matrix-synapse-py3, however, please read debian/NEWS before doing so. The existing matrix-synapse packages will continue to use Python 2 for the time being.

2. This release removes the riot.im from the default list of trusted identity servers.

   If riot.im is in your homeserver's list of trusted_third_party_id_servers, you should remove it. It was added in case a hypothetical future identity server was put there. If you don't remove it, users may be unable to deactivate their accounts.

3. This release no longer installs the (unmaintained) Matrix Console web client as part of the default installation. It is possible to re-enable it by installing it separately and setting the web_client_location config option, but please consider switching to another client.

Upgrading to v0.33.7

This release removes the example email notification templates from res/templates (they are now internal to the python package). This should only affect you if you (a) deploy your Synapse instance from a git checkout or a github snapshot URL, and (b) have email notifications enabled.

If you have email notifications enabled, you should ensure that email.template_dir is either configured to point at a directory where you have installed customised templates, or leave it unset to use the default templates.

Upgrading to v0.27.3

This release expands the anonymous usage stats sent if the opt-in report_stats configuration is set to true. We now capture RSS memory and cpu use at a very coarse level. This requires administrators to install the optional psutil python module.

We would appreciate it if you could assist by ensuring this module is available and report_stats is enabled. This will let us see if performance changes to synapse are having an impact to the general community.

Upgrading to v0.15.0

If you want to use the new URL previewing API (/_matrix/media/r0/preview_url) then you have to explicitly enable it in the config and update your dependencies dependencies. See README.rst for details.

Upgrading to v0.11.0

This release includes the option to send anonymous usage stats to matrix.org, and requires that administrators explictly opt in or out by setting the report_stats option to either true or false.

We would really appreciate it if you could help our project out by reporting anonymized usage statistics from your homeserver. Only very basic aggregate data (e.g. number of users) will be reported, but it helps us to track the growth of the Matrix community, and helps us to make Matrix a success, as well as to convince other networks that they should peer with us.

Upgrading to v0.9.0

Application services have had a breaking API change in this version.

They can no longer register themselves with a home server using the AS HTTP API. This decision was made because a compromised application service with free reign to register any regex in effect grants full read/write access to the home server if a regex of .* is used. An attack where a compromised AS re-registers itself with .* was deemed too big of a security risk to ignore, and so the ability to register with the HS remotely has been removed.

It has been replaced by specifying a list of application service registrations in homeserver.yaml:

app_service_config_files: ["registration-01.yaml", "registration-02.yaml"]

Where registration-01.yaml looks like:

url: <String>  # e.g. "https://my.application.service.com"
as_token: <String>
hs_token: <String>
sender_localpart: <String>  # This is a new field which denotes the user_id localpart when using the AS token
namespaces:
  users:
    - exclusive: <Boolean>
      regex: <String>  # e.g. "@prefix_.*"
  aliases:
    - exclusive: <Boolean>
      regex: <String>
  rooms:
    - exclusive: <Boolean>
      regex: <String>

Upgrading to v0.8.0

Servers which use captchas will need to add their public key to:

static/client/register/register_config.js

  window.matrixRegistrationConfig = {
      recaptcha_public_key: "YOUR_PUBLIC_KEY"
  };

This is required in order to support registration fallback (typically used on mobile devices).

Upgrading to v0.7.0

New dependencies are:

• pydenticon
• simplejson
• syutil
• matrix-angular-sdk

To pull in these dependencies in a virtual env, run:

python synapse/python_dependencies.py | xargs -n 1 pip install

Upgrading to v0.6.0

To pull in new dependencies, run:

python setup.py develop --user

This update includes a change to the database schema. To upgrade you first need to upgrade the database by running:

python scripts/upgrade_db_to_v0.6.0.py <db> <server_name> <signing_key>

Where <db> is the location of the database, <server_name> is the server name as specified in the synapse configuration, and <signing_key> is the location of the signing key as specified in the synapse configuration.

This may take some time to complete. Failures of signatures and content hashes can safely be ignored.

Upgrading to v0.5.1

Depending on precisely when you installed v0.5.0 you may have ended up with a stale release of the reference matrix webclient installed as a python module. To uninstall it and ensure you are depending on the latest module, please run:

$ pip uninstall syweb

Upgrading to v0.5.0

The webclient has been split out into a seperate repository/pacakage in this release. Before you restart your homeserver you will need to pull in the webclient package by running:

python setup.py develop --user

This release completely changes the database schema and so requires upgrading it before starting the new version of the homeserver.

The script "database-prepare-for-0.5.0.sh" should be used to upgrade the database. This will save all user information, such as logins and profiles, but will otherwise purge the database. This includes messages, which rooms the home server was a member of and room alias mappings.

If you would like to keep your history, please take a copy of your database file and ask for help in #matrix:matrix.org. The upgrade process is, unfortunately, non trivial and requires human intervention to resolve any resulting conflicts during the upgrade process.

Before running the command the homeserver should be first completely shutdown. To run it, simply specify the location of the database, e.g.:

./scripts/database-prepare-for-0.5.0.sh "homeserver.db"

Once this has successfully completed it will be safe to restart the homeserver. You may notice that the homeserver takes a few seconds longer to restart than usual as it reinitializes the database.

On startup of the new version, users can either rejoin remote rooms using room aliases or by being reinvited. Alternatively, if any other homeserver sends a message to a room that the homeserver was previously in the local HS will automatically rejoin the room.

Upgrading to v0.4.0

This release needs an updated syutil version. Run:

python setup.py develop

You will also need to upgrade your configuration as the signing key format has changed. Run:

python -m synapse.app.homeserver --config-path <CONFIG> --generate-config

Upgrading to v0.3.0

This registration API now closely matches the login API. This introduces a bit more backwards and forwards between the HS and the client, but this improves the overall flexibility of the API. You can now GET on /register to retrieve a list of valid registration flows. Upon choosing one, they are submitted in the same way as login, e.g:

{
  type: m.login.password,
  user: foo,
  password: bar
}

The default HS supports 2 flows, with and without Identity Server email authentication. Enabling captcha on the HS will add in an extra step to all flows: m.login.recaptcha which must be completed before you can transition to the next stage. There is a new login type: m.login.email.identity which contains the threepidCreds key which were previously sent in the original register request. For more information on this, see the specification.

Web Client

The VoIP specification has changed between v0.2.0 and v0.3.0. Users should refresh any browser tabs to get the latest web client code. Users on v0.2.0 of the web client will not be able to call those on v0.3.0 and vice versa.

Upgrading to v0.2.0

The home server now requires setting up of SSL config before it can run. To automatically generate default config use:

$ python synapse/app/homeserver.py \
    --server-name machine.my.domain.name \
    --bind-port 8448 \
    --config-path homeserver.config \
    --generate-config

This config can be edited if desired, for example to specify a different SSL certificate to use. Once done you can run the home server using:

$ python synapse/app/homeserver.py --config-path homeserver.config

See the README.rst for more information.

Also note that some config options have been renamed, including:

• "host" to "server-name"
• "database" to "database-path"
• "port" to "bind-port" and "unsecure-port"

Upgrading to v0.0.1

This release completely changes the database schema and so requires upgrading it before starting the new version of the homeserver.

The script "database-prepare-for-0.0.1.sh" should be used to upgrade the database. This will save all user information, such as logins and profiles, but will otherwise purge the database. This includes messages, which rooms the home server was a member of and room alias mappings.

Before running the command the homeserver should be first completely shutdown. To run it, simply specify the location of the database, e.g.:

./scripts/database-prepare-for-0.0.1.sh "homeserver.db"

Once this has successfully completed it will be safe to restart the homeserver. You may notice that the homeserver takes a few seconds longer to restart than usual as it reinitializes the database.

On startup of the new version, users can either rejoin remote rooms using room aliases or by being reinvited. Alternatively, if any other homeserver sends a message to a room that the homeserver was previously in the local HS will automatically rejoin the room.

Setting up federation

Federation is the process by which users on different servers can participate in the same room. For this to work, those other servers must be able to contact yours to send messages.

The server_name configured in the Synapse configuration file (often homeserver.yaml) defines how resources (users, rooms, etc.) will be identified (eg: @user:example.com, #room:example.com). By default, it is also the domain that other servers will use to try to reach your server (via port 8448). This is easy to set up and will work provided you set the server_name to match your machine's public DNS hostname.

For this default configuration to work, you will need to listen for TLS connections on port 8448. The preferred way to do that is by using a reverse proxy: see the reverse proxy documentation for instructions on how to correctly set one up.

In some cases you might not want to run Synapse on the machine that has the server_name as its public DNS hostname, or you might want federation traffic to use a different port than 8448. For example, you might want to have your user names look like @user:example.com, but you want to run Synapse on synapse.example.com on port 443. This can be done using delegation, which allows an admin to control where federation traffic should be sent. See the delegation documentation for instructions on how to set this up.

Once federation has been configured, you should be able to join a room over federation. A good place to start is #synapse:matrix.org - a room for Synapse admins.

Troubleshooting

You can use the federation tester to check if your homeserver is configured correctly. Alternatively try the JSON API used by the federation tester. Note that you'll have to modify this URL to replace DOMAIN with your server_name. Hitting the API directly provides extra detail.

The typical failure mode for federation is that when the server tries to join a room, it is rejected with "401: Unauthorized". Generally this means that other servers in the room could not access yours. (Joining a room over federation is a complicated dance which requires connections in both directions).

Another common problem is that people on other servers can't join rooms that you invite them to. This can be caused by an incorrectly-configured reverse proxy: see the reverse proxy documentation for instructions on how to correctly configure a reverse proxy.

Known issues

HTTP 308 Permanent Redirect redirects are not followed: Due to missing features in the HTTP library used by Synapse, 308 redirects are currently not followed by federating servers, which can cause M_UNKNOWN or 401 Unauthorized errors. This may affect users who are redirecting apex-to-www (e.g. example.com -> www.example.com), and especially users of the Kubernetes Nginx Ingress module, which uses 308 redirect codes by default. For those Kubernetes users, this Stackoverflow post might be helpful. For other users, switching to a 301 Moved Permanently code may be an option. 308 redirect codes will be supported properly in a future release of Synapse.

Running a demo federation of Synapses

If you want to get up and running quickly with a trio of homeservers in a private federation, there is a script in the demo directory. This is mainly useful just for development purposes. See demo scripts.

Configuration

This section contains information on tweaking Synapse via the various options in the configuration file. A configuration file should have been generated when you installed Synapse.

Configuring Synapse

This is intended as a guide to the Synapse configuration. The behavior of a Synapse instance can be modified through the many configuration settings documented here — each config option is explained, including what the default is, how to change the default and what sort of behaviour the setting governs. Also included is an example configuration for each setting. If you don't want to spend a lot of time thinking about options, the config as generated sets sensible defaults for all values. Do note however that the database defaults to SQLite, which is not recommended for production usage. You can read more on this subject here.

Config Conventions

Configuration options that take a time period can be set using a number followed by a letter. Letters have the following meanings:

• s = second
• m = minute
• h = hour
• d = day
• w = week
• y = year

For example, setting redaction_retention_period: 5m would remove redacted messages from the database after 5 minutes, rather than 5 months.

In addition, configuration options referring to size use the following suffixes:

• M = MiB, or 1,048,576 bytes
• K = KiB, or 1024 bytes

For example, setting max_avatar_size: 10M means that Synapse will not accept files larger than 10,485,760 bytes for a user avatar.

YAML

The configuration file is a YAML file, which means that certain syntax rules apply if you want your config file to be read properly. A few helpful things to know:

• # before any option in the config will comment out that setting and either a default (if available) will be applied or Synapse will ignore the setting. Thus, in example #1 below, the setting will be read and applied, but in example #2 the setting will not be read and a default will be applied.

  Example #1:

  pid_file: DATADIR/homeserver.pid
  

  Example #2:

  #pid_file: DATADIR/homeserver.pid
  

• Indentation matters! The indentation before a setting will determine whether a given setting is read as part of another setting, or considered on its own. Thus, in example #1, the enabled setting is read as a sub-option of the presence setting, and will be properly applied.

  However, the lack of indentation before the enabled setting in example #2 means that when reading the config, Synapse will consider both presence and enabled as different settings. In this case, presence has no value, and thus a default applied, and enabled is an option that Synapse doesn't recognize and thus ignores.

  Example #1:

  presence:
    enabled: false
  

  Example #2:

  presence:
  enabled: false
  

  In this manual, all top-level settings (ones with no indentation) are identified at the beginning of their section (i.e. "### example_setting") and the sub-options, if any, are identified and listed in the body of the section. In addition, each setting has an example of its usage, with the proper indentation shown.

Modules

Server admins can expand Synapse's functionality with external modules.

See here for more documentation on how to configure or create custom modules for Synapse.

modules

Use the module sub-option to add modules under this option to extend functionality. The module setting then has a sub-option, config, which can be used to define some configuration for the module.

Defaults to none.

Example configuration:

modules:
  - module: my_super_module.MySuperClass
    config:
      do_thing: true
  - module: my_other_super_module.SomeClass
    config: {}

Server

Define your homeserver name and other base options.

server_name

This sets the public-facing domain of the server.

The server_name name will appear at the end of usernames and room addresses created on your server. For example if the server_name was example.com, usernames on your server would be in the format @user:example.com

In most cases you should avoid using a matrix specific subdomain such as matrix.example.com or synapse.example.com as the server_name for the same reasons you wouldn't use user@email.example.com as your email address. See here for information on how to host Synapse on a subdomain while preserving a clean server_name.

The server_name cannot be changed later so it is important to configure this correctly before you start Synapse. It should be all lowercase and may contain an explicit port.

There is no default for this option.

Example configuration #1:

server_name: matrix.org

Example configuration #2:

server_name: localhost:8080

pid_file

When running Synapse as a daemon, the file to store the pid in. Defaults to none.

Example configuration:

pid_file: DATADIR/homeserver.pid

web_client_location

The absolute URL to the web client which / will redirect to. Defaults to none.

Example configuration:

web_client_location: https://riot.example.com/

public_baseurl

The public-facing base URL that clients use to access this Homeserver (not including _matrix/...). This is the same URL a user might enter into the 'Custom Homeserver URL' field on their client. If you use Synapse with a reverse proxy, this should be the URL to reach Synapse via the proxy. Otherwise, it should be the URL to reach Synapse's client HTTP listener (see 'listeners' below).

Defaults to https://<server_name>/.

Example configuration:

public_baseurl: https://example.com/

serve_server_wellknown

By default, other servers will try to reach our server on port 8448, which can be inconvenient in some environments.

Provided https://<server_name>/ on port 443 is routed to Synapse, this option configures Synapse to serve a file at https://<server_name>/.well-known/matrix/server. This will tell other servers to send traffic to port 443 instead.

This option currently defaults to false.

See https://matrix-org.github.io/synapse/latest/delegate.html for more information.

Example configuration:

serve_server_wellknown: true

extra_well_known_client_content

This option allows server runners to add arbitrary key-value pairs to the client-facing .well-known response. Note that the public_baseurl config option must be provided for Synapse to serve a response to /.well-known/matrix/client at all.

If this option is provided, it parses the given yaml to json and serves it on /.well-known/matrix/client endpoint alongside the standard properties.

Added in Synapse 1.62.0.

Example configuration:

extra_well_known_client_content :
  option1: value1
  option2: value2

soft_file_limit

Set the soft limit on the number of file descriptors synapse can use. Zero is used to indicate synapse should set the soft limit to the hard limit. Defaults to 0.

Example configuration:

soft_file_limit: 3

presence

Presence tracking allows users to see the state (e.g online/offline) of other local and remote users. Set the enabled sub-option to false to disable presence tracking on this homeserver. Defaults to true. This option replaces the previous top-level 'use_presence' option.

Example configuration:

presence:
  enabled: false

require_auth_for_profile_requests

Whether to require authentication to retrieve profile data (avatars, display names) of other users through the client API. Defaults to false. Note that profile data is also available via the federation API, unless allow_profile_lookup_over_federation is set to false.

Example configuration:

require_auth_for_profile_requests: true

limit_profile_requests_to_users_who_share_rooms

Use this option to require a user to share a room with another user in order to retrieve their profile information. Only checked on Client-Server requests. Profile requests from other servers should be checked by the requesting server. Defaults to false.

Example configuration:

limit_profile_requests_to_users_who_share_rooms: true

include_profile_data_on_invite

Use this option to prevent a user's profile data from being retrieved and displayed in a room until they have joined it. By default, a user's profile data is included in an invite event, regardless of the values of the above two settings, and whether or not the users share a server. Defaults to true.

Example configuration:

include_profile_data_on_invite: false

allow_public_rooms_without_auth

If set to true, removes the need for authentication to access the server's public rooms directory through the client API, meaning that anyone can query the room directory. Defaults to false.

Example configuration:

allow_public_rooms_without_auth: true

allow_public_rooms_over_federation

If set to true, allows any other homeserver to fetch the server's public rooms directory via federation. Defaults to false.

Example configuration:

allow_public_rooms_over_federation: true

default_room_version

The default room version for newly created rooms on this server.

Known room versions are listed here

For example, for room version 1, default_room_version should be set to "1".

Currently defaults to "9".

Example configuration:

default_room_version: "8"

gc_thresholds

The garbage collection threshold parameters to pass to gc.set_threshold, if defined. Defaults to none.

Example configuration:

gc_thresholds: [700, 10, 10]

gc_min_interval

The minimum time in seconds between each GC for a generation, regardless of the GC thresholds. This ensures that we don't do GC too frequently. A value of [1s, 10s, 30s] indicates that a second must pass between consecutive generation 0 GCs, etc.

Defaults to [1s, 10s, 30s].

Example configuration:

gc_min_interval: [0.5s, 30s, 1m]

filter_timeline_limit

Set the limit on the returned events in the timeline in the get and sync operations. Defaults to 100. A value of -1 means no upper limit.

Example configuration:

filter_timeline_limit: 5000

block_non_admin_invites

Whether room invites to users on this server should be blocked (except those sent by local server admins). Defaults to false.

Example configuration:

block_non_admin_invites: true

enable_search

If set to false, new messages will not be indexed for searching and users will receive errors when searching for messages. Defaults to true.

Example configuration:

enable_search: false

ip_range_blacklist

This option prevents outgoing requests from being sent to the specified blacklisted IP address CIDR ranges. If this option is not specified then it defaults to private IP address ranges (see the example below).

The blacklist applies to the outbound requests for federation, identity servers, push servers, and for checking key validity for third-party invite events.

(0.0.0.0 and :: are always blacklisted, whether or not they are explicitly listed here, since they correspond to unroutable addresses.)

This option replaces federation_ip_range_blacklist in Synapse v1.25.0.

Note: The value is ignored when an HTTP proxy is in use.

Example configuration:

ip_range_blacklist:
  - '127.0.0.0/8'
  - '10.0.0.0/8'
  - '172.16.0.0/12'
  - '192.168.0.0/16'
  - '100.64.0.0/10'
  - '192.0.0.0/24'
  - '169.254.0.0/16'
  - '192.88.99.0/24'
  - '198.18.0.0/15'
  - '192.0.2.0/24'
  - '198.51.100.0/24'
  - '203.0.113.0/24'
  - '224.0.0.0/4'
  - '::1/128'
  - 'fe80::/10'
  - 'fc00::/7'
  - '2001:db8::/32'
  - 'ff00::/8'
  - 'fec0::/10'

ip_range_whitelist

List of IP address CIDR ranges that should be allowed for federation, identity servers, push servers, and for checking key validity for third-party invite events. This is useful for specifying exceptions to wide-ranging blacklisted target IP ranges - e.g. for communication with a push server only visible in your network.

This whitelist overrides ip_range_blacklist and defaults to an empty list.

Example configuration:

ip_range_whitelist:
   - '192.168.1.1'

listeners

List of ports that Synapse should listen on, their purpose and their configuration.

Sub-options for each listener include:

• port: the TCP port to bind to.

• bind_addresses: a list of local addresses to listen on. The default is 'all local interfaces'.

• type: the type of listener. Normally http, but other valid options are:

  • manhole: (see the docs here),

  • metrics: (see the docs here),

• tls: set to true to enable TLS for this listener. Will use the TLS key/cert specified in tls_private_key_path / tls_certificate_path.

• x_forwarded: Only valid for an 'http' listener. Set to true to use the X-Forwarded-For header as the client IP. Useful when Synapse is behind a reverse-proxy.

• resources: Only valid for an 'http' listener. A list of resources to host on this port. Sub-options for each resource are:

  • names: a list of names of HTTP resources. See below for a list of valid resource names.

  • compress: set to true to enable gzip compression on HTTP bodies for this resource. This is currently only supported with the client, consent, metrics and federation resources.

• additional_resources: Only valid for an 'http' listener. A map of additional endpoints which should be loaded via dynamic modules.

Valid resource names are:

• client: the client-server API (/_matrix/client), and the synapse admin API (/_synapse/admin). Also implies media and static.

• consent: user consent forms (/_matrix/consent). See here for more.

• federation: the server-server API (/_matrix/federation). Also implies media, keys, openid

• keys: the key discovery API (/_matrix/key).

• media: the media API (/_matrix/media).

• metrics: the metrics interface. See here.

• openid: OpenID authentication. See here.

• replication: the HTTP replication API (/_synapse/replication). See here.

• static: static resources under synapse/static (/_matrix/static). (Mostly useful for 'fallback authentication'.)

Example configuration #1:

listeners:
  # TLS-enabled listener: for when matrix traffic is sent directly to synapse.
  #
  # (Note that you will also need to give Synapse a TLS key and certificate: see the TLS section
  # below.)
  #
  - port: 8448
    type: http
    tls: true
    resources:
      - names: [client, federation]

Example configuration #2:

listeners:
  # Unsecure HTTP listener: for when matrix traffic passes through a reverse proxy
  # that unwraps TLS.
  #
  # If you plan to use a reverse proxy, please see
  # https://matrix-org.github.io/synapse/latest/reverse_proxy.html.
  #
  - port: 8008
    tls: false
    type: http
    x_forwarded: true
    bind_addresses: ['::1', '127.0.0.1']

    resources:
      - names: [client, federation]
        compress: false

    # example additional_resources:
    additional_resources:
      "/_matrix/my/custom/endpoint":
        module: my_module.CustomRequestHandler
        config: {}

  # Turn on the twisted ssh manhole service on localhost on the given
  # port.
  - port: 9000
    bind_addresses: ['::1', '127.0.0.1']
    type: manhole

manhole_settings

Connection settings for the manhole. You can find more information on the manhole here. Manhole sub-options include:

• username : the username for the manhole. This defaults to 'matrix'.
• password: The password for the manhole. This defaults to 'rabbithole'.
• ssh_priv_key_path and ssh_pub_key_path: The private and public SSH key pair used to encrypt the manhole traffic. If these are left unset, then hardcoded and non-secret keys are used, which could allow traffic to be intercepted if sent over a public network.

Example configuration:

manhole_settings:
  username: manhole
  password: mypassword
  ssh_priv_key_path: CONFDIR/id_rsa
  ssh_pub_key_path: CONFDIR/id_rsa.pub

dummy_events_threshold

Forward extremities can build up in a room due to networking delays between homeservers. Once this happens in a large room, calculation of the state of that room can become quite expensive. To mitigate this, once the number of forward extremities reaches a given threshold, Synapse will send an org.matrix.dummy_event event, which will reduce the forward extremities in the room.

This setting defines the threshold (i.e. number of forward extremities in the room) at which dummy events are sent. The default value is 10.

Example configuration:

dummy_events_threshold: 5

delete_stale_devices_after

An optional duration. If set, Synapse will run a daily background task to log out and delete any device that hasn't been accessed for more than the specified amount of time.

Defaults to no duration, which means devices are never pruned.

Example configuration:

delete_stale_devices_after: 1y

Homeserver blocking

Useful options for Synapse admins.

admin_contact

How to reach the server admin, used in ResourceLimitError. Defaults to none.

Example configuration:

admin_contact: 'mailto:admin@server.com'

hs_disabled and hs_disabled_message

Blocks users from connecting to the homeserver and provides a human-readable reason why the connection was blocked. Defaults to false.

Example configuration:

hs_disabled: true
hs_disabled_message: 'Reason for why the HS is blocked'

limit_usage_by_mau

This option disables/enables monthly active user blocking. Used in cases where the admin or server owner wants to limit to the number of monthly active users. When enabled and a limit is reached the server returns a ResourceLimitError with error type Codes.RESOURCE_LIMIT_EXCEEDED. Defaults to false. If this is enabled, a value for max_mau_value must also be set.

See Monthly Active Users for details on how to configure MAU.

Example configuration:

limit_usage_by_mau: true

max_mau_value

This option sets the hard limit of monthly active users above which the server will start blocking user actions if limit_usage_by_mau is enabled. Defaults to 0.

Example configuration:

max_mau_value: 50

mau_trial_days

The option mau_trial_days is a means to add a grace period for active users. It means that users must be active for the specified number of days before they can be considered active and guards against the case where lots of users sign up in a short space of time never to return after their initial session. Defaults to 0.

Example configuration:

mau_trial_days: 5

mau_appservice_trial_days

The option mau_appservice_trial_days is similar to mau_trial_days, but applies a different trial number if the user was registered by an appservice. A value of 0 means no trial days are applied. Appservices not listed in this dictionary use the value of mau_trial_days instead.

Example configuration:

mau_appservice_trial_days:
  my_appservice_id: 3
  another_appservice_id: 6

mau_limit_alerting

The option mau_limit_alerting is a means of limiting client-side alerting should the mau limit be reached. This is useful for small instances where the admin has 5 mau seats (say) for 5 specific people and no interest increasing the mau limit further. Defaults to true, which means that alerting is enabled.

Example configuration:

mau_limit_alerting: false

mau_stats_only

If enabled, the metrics for the number of monthly active users will be populated, however no one will be limited based on these numbers. If limit_usage_by_mau is true, this is implied to be true. Defaults to false.

Example configuration:

mau_stats_only: true

mau_limit_reserved_threepids

Sometimes the server admin will want to ensure certain accounts are never blocked by mau checking. These accounts are specified by this option. Defaults to none. Add accounts by specifying the medium and address of the reserved threepid (3rd party identifier).

Example configuration:

mau_limit_reserved_threepids:
  - medium: 'email'
    address: 'reserved_user@example.com'

server_context

This option is used by phonehome stats to group together related servers. Defaults to none.

Example configuration:

server_context: context

limit_remote_rooms

When this option is enabled, the room "complexity" will be checked before a user joins a new remote room. If it is above the complexity limit, the server will disallow joining, or will instantly leave. This is useful for homeservers that are resource-constrained. Options for this setting include:

• enabled: whether this check is enabled. Defaults to false.
• complexity: the limit above which rooms cannot be joined. The default is 1.0.
• complexity_error: override the error which is returned when the room is too complex with a custom message.
• admins_can_join: allow server admins to join complex rooms. Default is false.

Room complexity is an arbitrary measure based on factors such as the number of users in the room.

Example configuration:

limit_remote_rooms:
  enabled: true
  complexity: 0.5
  complexity_error: "I can't let you do that, Dave."
  admins_can_join: true

require_membership_for_aliases

Whether to require a user to be in the room to add an alias to it. Defaults to true.

Example configuration:

require_membership_for_aliases: false

allow_per_room_profiles

Whether to allow per-room membership profiles through the sending of membership events with profile information that differs from the target's global profile. Defaults to true.

Example configuration:

allow_per_room_profiles: false

max_avatar_size

The largest permissible file size in bytes for a user avatar. Defaults to no restriction. Use M for MB and K for KB.

Note that user avatar changes will not work if this is set without using Synapse's media repository.

Example configuration:

max_avatar_size: 10M

allowed_avatar_mimetypes

The MIME types allowed for user avatars. Defaults to no restriction.

Note that user avatar changes will not work if this is set without using Synapse's media repository.

Example configuration:

allowed_avatar_mimetypes: ["image/png", "image/jpeg", "image/gif"]

redaction_retention_period

How long to keep redacted events in unredacted form in the database. After this period redacted events get replaced with their redacted form in the DB.

Synapse will check whether the rentention period has concluded for redacted events every 5 minutes. Thus, even if this option is set to 0, Synapse may still take up to 5 minutes to purge redacted events from the database.

Defaults to 7d. Set to null to disable.

Example configuration:

redaction_retention_period: 28d

user_ips_max_age

How long to track users' last seen time and IPs in the database.

Defaults to 28d. Set to null to disable clearing out of old rows.

Example configuration:

user_ips_max_age: 14d

request_token_inhibit_3pid_errors

Inhibits the /requestToken endpoints from returning an error that might leak information about whether an e-mail address is in use or not on this homeserver. Defaults to false. Note that for some endpoints the error situation is the e-mail already being used, and for others the error is entering the e-mail being unused. If this option is enabled, instead of returning an error, these endpoints will act as if no error happened and return a fake session ID ('sid') to clients.

Example configuration:

request_token_inhibit_3pid_errors: true

next_link_domain_whitelist

A list of domains that the domain portion of next_link parameters must match.

This parameter is optionally provided by clients while requesting validation of an email or phone number, and maps to a link that users will be automatically redirected to after validation succeeds. Clients can make use this parameter to aid the validation process.

The whitelist is applied whether the homeserver or an identity server is handling validation.

The default value is no whitelist functionality; all domains are allowed. Setting this value to an empty list will instead disallow all domains.

Example configuration:

next_link_domain_whitelist: ["matrix.org"]

templates and custom_template_directory

These options define templates to use when generating email or HTML page contents. The custom_template_directory determines which directory Synapse will try to find template files in to use to generate email or HTML page contents. If not set, or a file is not found within the template directory, a default template from within the Synapse package will be used.

See here for more information about using custom templates.

Example configuration:

templates:
  custom_template_directory: /path/to/custom/templates/

retention

This option and the associated options determine message retention policy at the server level.

Room admins and mods can define a retention period for their rooms using the m.room.retention state event, and server admins can cap this period by setting the allowed_lifetime_min and allowed_lifetime_max config options.

If this feature is enabled, Synapse will regularly look for and purge events which are older than the room's maximum retention period. Synapse will also filter events received over federation so that events that should have been purged are ignored and not stored again.

The message retention policies feature is disabled by default. Please be advised that enabling this feature carries some risk. There are known bugs with the implementation which can cause database corruption. Setting retention to delete older history is less risky than deleting newer history but in general caution is advised when enabling this experimental feature. You can read more about this feature here.

This setting has the following sub-options:

• default_policy: Default retention policy. If set, Synapse will apply it to rooms that lack the 'm.room.retention' state event. This option is further specified by the min_lifetime and max_lifetime sub-options associated with it. Note that the value of min_lifetime doesn't matter much because Synapse doesn't take it into account yet.

• allowed_lifetime_min and allowed_lifetime_max: Retention policy limits. If set, and the state of a room contains a m.room.retention event in its state which contains a min_lifetime or a max_lifetime that's out of these bounds, Synapse will cap the room's policy to these limits when running purge jobs.

• purge_jobs and the associated shortest_max_lifetime and longest_max_lifetime sub-options: Server admins can define the settings of the background jobs purging the events whose lifetime has expired under the purge_jobs section.

  If no configuration is provided for this option, a single job will be set up to delete expired events in every room daily.

  Each job's configuration defines which range of message lifetimes the job takes care of. For example, if shortest_max_lifetime is '2d' and longest_max_lifetime is '3d', the job will handle purging expired events in rooms whose state defines a max_lifetime that's both higher than 2 days, and lower than or equal to 3 days. Both the minimum and the maximum value of a range are optional, e.g. a job with no shortest_max_lifetime and a longest_max_lifetime of '3d' will handle every room with a retention policy whose max_lifetime is lower than or equal to three days.

  The rationale for this per-job configuration is that some rooms might have a retention policy with a low max_lifetime, where history needs to be purged of outdated messages on a more frequent basis than for the rest of the rooms (e.g. every 12h), but not want that purge to be performed by a job that's iterating over every room it knows, which could be heavy on the server.

  If any purge job is configured, it is strongly recommended to have at least a single job with neither shortest_max_lifetime nor longest_max_lifetime set, or one job without shortest_max_lifetime and one job without longest_max_lifetime set. Otherwise some rooms might be ignored, even if allowed_lifetime_min and allowed_lifetime_max are set, because capping a room's policy to these values is done after the policies are retrieved from Synapse's database (which is done using the range specified in a purge job's configuration).

Example configuration:

retention:
  enabled: true
  default_policy:
    min_lifetime: 1d
    max_lifetime: 1y
  allowed_lifetime_min: 1d
  allowed_lifetime_max: 1y
  purge_jobs:
    - longest_max_lifetime: 3d
      interval: 12h
    - shortest_max_lifetime: 3d
      interval: 1d

TLS

Options related to TLS.

tls_certificate_path

This option specifies a PEM-encoded X509 certificate for TLS. This certificate, as of Synapse 1.0, will need to be a valid and verifiable certificate, signed by a recognised Certificate Authority. Defaults to none.

Be sure to use a .pem file that includes the full certificate chain including any intermediate certificates (for instance, if using certbot, use fullchain.pem as your certificate, not cert.pem).

Example configuration:

tls_certificate_path: "CONFDIR/SERVERNAME.tls.crt"

tls_private_key_path

PEM-encoded private key for TLS. Defaults to none.

Example configuration:

tls_private_key_path: "CONFDIR/SERVERNAME.tls.key"

federation_verify_certificates

Whether to verify TLS server certificates for outbound federation requests.

Defaults to true. To disable certificate verification, set the option to false.

Example configuration:

federation_verify_certificates: false

federation_client_minimum_tls_version

The minimum TLS version that will be used for outbound federation requests.

Defaults to 1. Configurable to 1, 1.1, 1.2, or 1.3. Note that setting this value higher than 1.2 will prevent federation to most of the public Matrix network: only configure it to 1.3 if you have an entirely private federation setup and you can ensure TLS 1.3 support.

Example configuration:

federation_client_minimum_tls_version: 1.2

federation_certificate_verification_whitelist

Skip federation certificate verification on a given whitelist of domains.

This setting should only be used in very specific cases, such as federation over Tor hidden services and similar. For private networks of homeservers, you likely want to use a private CA instead.

Only effective if federation_verify_certicates is true.

Example configuration:

federation_certificate_verification_whitelist:
  - lon.example.com
  - "*.domain.com"
  - "*.onion"

federation_custom_ca_list

List of custom certificate authorities for federation traffic.

This setting should only normally be used within a private network of homeservers.

Note that this list will replace those that are provided by your operating environment. Certificates must be in PEM format.

Example configuration:

federation_custom_ca_list:
  - myCA1.pem
  - myCA2.pem
  - myCA3.pem

Federation

Options related to federation.

federation_domain_whitelist

Restrict federation to the given whitelist of domains. N.B. we recommend also firewalling your federation listener to limit inbound federation traffic as early as possible, rather than relying purely on this application-layer restriction. If not specified, the default is to whitelist everything.

Example configuration:

federation_domain_whitelist:
  - lon.example.com
  - nyc.example.com
  - syd.example.com

federation_metrics_domains

Report prometheus metrics on the age of PDUs being sent to and received from the given domains. This can be used to give an idea of "delay" on inbound and outbound federation, though be aware that any delay can be due to problems at either end or with the intermediate network.

By default, no domains are monitored in this way.

Example configuration:

federation_metrics_domains:
  - matrix.org
  - example.com

allow_profile_lookup_over_federation

Set to false to disable profile lookup over federation. By default, the Federation API allows other homeservers to obtain profile data of any user on this homeserver.

Example configuration:

allow_profile_lookup_over_federation: false

allow_device_name_lookup_over_federation

Set this option to true to allow device display name lookup over federation. By default, the Federation API prevents other homeservers from obtaining the display names of any user devices on this homeserver.

Example configuration:

allow_device_name_lookup_over_federation: true

Caching

Options related to caching.

event_cache_size

The number of events to cache in memory. Defaults to 10K. Like other caches, this is affected by caches.global_factor (see below).

Note that this option is not part of the caches section.

Example configuration:

event_cache_size: 15K

caches and associated values

A cache 'factor' is a multiplier that can be applied to each of Synapse's caches in order to increase or decrease the maximum number of entries that can be stored.

caches can be configured through the following sub-options:

• global_factor: Controls the global cache factor, which is the default cache factor for all caches if a specific factor for that cache is not otherwise set.

  This can also be set by the SYNAPSE_CACHE_FACTOR environment variable. Setting by environment variable takes priority over setting through the config file.

  Defaults to 0.5, which will halve the size of all caches.

• per_cache_factors: A dictionary of cache name to cache factor for that individual cache. Overrides the global cache factor for a given cache.

  These can also be set through environment variables comprised of SYNAPSE_CACHE_FACTOR_ + the name of the cache in capital letters and underscores. Setting by environment variable takes priority over setting through the config file. Ex. SYNAPSE_CACHE_FACTOR_GET_USERS_WHO_SHARE_ROOM_WITH_USER=2.0

  Some caches have '*' and other characters that are not alphanumeric or underscores. These caches can be named with or without the special characters stripped. For example, to specify the cache factor for *stateGroupCache* via an environment variable would be SYNAPSE_CACHE_FACTOR_STATEGROUPCACHE=2.0.

• expire_caches: Controls whether cache entries are evicted after a specified time period. Defaults to true. Set to false to disable this feature. Note that never expiring caches may result in excessive memory usage.

• cache_entry_ttl: If expire_caches is enabled, this flag controls how long an entry can be in a cache without having been accessed before being evicted. Defaults to 30m.

• sync_response_cache_duration: Controls how long the results of a /sync request are cached for after a successful response is returned. A higher duration can help clients with intermittent connections, at the cost of higher memory usage. A value of zero means that sync responses are not cached. Defaults to 2m.

  Changed in Synapse 1.62.0: The default was changed from 0 to 2m.

• cache_autotuning and its sub-options max_cache_memory_usage, target_cache_memory_usage, and min_cache_ttl work in conjunction with each other to maintain a balance between cache memory usage and cache entry availability. You must be using jemalloc to utilize this option, and all three of the options must be specified for this feature to work. This option defaults to off, enable it by providing values for the sub-options listed below. Please note that the feature will not work and may cause unstable behavior (such as excessive emptying of caches or exceptions) if all of the values are not provided. Please see the Config Conventions for information on how to specify memory size and cache expiry durations.

  • max_cache_memory_usage sets a ceiling on how much memory the cache can use before caches begin to be continuously evicted. They will continue to be evicted until the memory usage drops below the target_memory_usage, set in the setting below, or until the min_cache_ttl is hit. There is no default value for this option.
  • target_memory_usage sets a rough target for the desired memory usage of the caches. There is no default value for this option.
  • min_cache_ttl sets a limit under which newer cache entries are not evicted and is only applied when caches are actively being evicted/max_cache_memory_usage has been exceeded. This is to protect hot caches from being emptied while Synapse is evicting due to memory. There is no default value for this option.

Example configuration:

event_cache_size: 15K
caches:
  global_factor: 1.0
  per_cache_factors:
    get_users_who_share_room_with_user: 2.0
  sync_response_cache_duration: 2m
  cache_autotuning:
    max_cache_memory_usage: 1024M
    target_cache_memory_usage: 758M
    min_cache_ttl: 5m

Reloading cache factors

The cache factors (i.e. caches.global_factor and caches.per_cache_factors) may be reloaded at any time by sending a SIGHUP signal to Synapse using e.g.

kill -HUP [PID_OF_SYNAPSE_PROCESS]

If you are running multiple workers, you must individually update the worker config file and send this signal to each worker process.

If you're using the example systemd service file in Synapse's contrib directory, you can send a SIGHUP signal by using systemctl reload matrix-synapse.

Database

Config options related to database settings.

database

The database setting defines the database that synapse uses to store all of its data.

Associated sub-options:

• name: this option specifies the database engine to use: either sqlite3 (for SQLite) or psycopg2 (for PostgreSQL). If no name is specified Synapse will default to SQLite.

• txn_limit gives the maximum number of transactions to run per connection before reconnecting. Defaults to 0, which means no limit.

• allow_unsafe_locale is an option specific to Postgres. Under the default behavior, Synapse will refuse to start if the postgres db is set to a non-C locale. You can override this behavior (which is not recommended) by setting allow_unsafe_locale to true. Note that doing so may corrupt your database. You can find more information here and here.

• args gives options which are passed through to the database engine, except for options starting with cp_, which are used to configure the Twisted connection pool. For a reference to valid arguments, see:

  • for sqlite
  • for postgres
  • for the connection pool

For more information on using Synapse with Postgres, see here.

Example SQLite configuration:

database:
  name: sqlite3
  args:
    database: /path/to/homeserver.db

Example Postgres configuration:

database:
  name: psycopg2
  txn_limit: 10000
  args:
    user: synapse_user
    password: secretpassword
    database: synapse
    host: localhost
    port: 5432
    cp_min: 5
    cp_max: 10

databases

The databases option allows specifying a mapping between certain database tables and database host details, spreading the load of a single Synapse instance across multiple database backends. This is often referred to as "database sharding". This option is only supported for PostgreSQL database backends.

Important note: This is a supported option, but is not currently used in production by the Matrix.org Foundation. Proceed with caution and always make backups.

databases is a dictionary of arbitrarily-named database entries. Each entry is equivalent to the value of the database homeserver config option (see above), with the addition of a data_stores key. data_stores is an array of strings that specifies the data store(s) (a defined label for a set of tables) that should be stored on the associated database backend entry.

The currently defined values for data_stores are:

• "state": Database that relates to state groups will be stored in this database.

  Specifically, that means the following tables:

  • state_groups
  • state_group_edges
  • state_groups_state

  And the following sequences:

  • state_groups_seq_id

• "main": All other database tables and sequences.

All databases will end up with additional tables used for tracking database schema migrations and any pending background updates. Synapse will create these automatically on startup when checking for and/or performing database schema migrations.

To migrate an existing database configuration (e.g. all tables on a single database) to a different configuration (e.g. the "main" data store on one database, and "state" on another), do the following:

1. Take a backup of your existing database. Things can and do go wrong and database corruption is no joke!

2. Ensure all pending database migrations have been applied and background updates have run. The simplest way to do this is to use the update_synapse_database script supplied with your Synapse installation.

   update_synapse_database --database-config homeserver.yaml --run-background-updates
   

3. Copy over the necessary tables and sequences from one database to the other. Tables relating to database migrations, schemas, schema versions and background updates should not be copied.

   As an example, say that you'd like to split out the "state" data store from an existing database which currently contains all data stores.

   Simply copy the tables and sequences defined above for the "state" datastore from the existing database to the secondary database. As noted above, additional tables will be created in the secondary database when Synapse is started.

4. Modify/create the databases option in your homeserver.yaml to match the desired database configuration.

5. Start Synapse. Check that it starts up successfully and that things generally seem to be working.

6. Drop the old tables that were copied in step 3.

Only one of the options database or databases may be specified in your config, but not both.

Example configuration:

databases:
  basement_box:
    name: psycopg2
    txn_limit: 10000
    data_stores: ["main"]
    args:
      user: synapse_user
      password: secretpassword
      database: synapse_main
      host: localhost
      port: 5432
      cp_min: 5
      cp_max: 10

  my_other_database:
    name: psycopg2
    txn_limit: 10000
    data_stores: ["state"]
    args:
      user: synapse_user
      password: secretpassword
      database: synapse_state
      host: localhost
      port: 5432
      cp_min: 5
      cp_max: 10

Logging

Config options related to logging.

log_config

This option specifies a yaml python logging config file as described here.

Example configuration:

log_config: "CONFDIR/SERVERNAME.log.config"

Ratelimiting

Options related to ratelimiting in Synapse.

Each ratelimiting configuration is made of two parameters:

• per_second: number of requests a client can send per second.
• burst_count: number of requests a client can send before being throttled.

rc_message

Ratelimiting settings for client messaging.

This is a ratelimiting option for messages that ratelimits sending based on the account the client is using. It defaults to: per_second: 0.2, burst_count: 10.

Example configuration:

rc_message:
  per_second: 0.5
  burst_count: 15

rc_registration

This option ratelimits registration requests based on the client's IP address. It defaults to per_second: 0.17, burst_count: 3.

Example configuration:

rc_registration:
  per_second: 0.15
  burst_count: 2

rc_registration_token_validity

This option checks the validity of registration tokens that ratelimits requests based on the client's IP address. Defaults to per_second: 0.1, burst_count: 5.

Example configuration:

rc_registration_token_validity:
  per_second: 0.3
  burst_count: 6

rc_login

This option specifies several limits for login:

• address ratelimits login requests based on the client's IP address. Defaults to per_second: 0.17, burst_count: 3.

• account ratelimits login requests based on the account the client is attempting to log into. Defaults to per_second: 0.17, burst_count: 3.

• failed_attempts ratelimits login requests based on the account the client is attempting to log into, based on the amount of failed login attempts for this account. Defaults to per_second: 0.17, burst_count: 3.

Example configuration:

rc_login:
  address:
    per_second: 0.15
    burst_count: 5
  account:
    per_second: 0.18
    burst_count: 4
  failed_attempts:
    per_second: 0.19
    burst_count: 7

rc_admin_redaction

This option sets ratelimiting redactions by room admins. If this is not explicitly set then it uses the same ratelimiting as per rc_message. This is useful to allow room admins to deal with abuse quickly.

Example configuration:

rc_admin_redaction:
  per_second: 1
  burst_count: 50

rc_joins

This option allows for ratelimiting number of rooms a user can join. This setting has the following sub-options:

• local: ratelimits when users are joining rooms the server is already in. Defaults to per_second: 0.1, burst_count: 10.

• remote: ratelimits when users are trying to join rooms not on the server (which can be more computationally expensive than restricting locally). Defaults to per_second: 0.01, burst_count: 10

Example configuration:

rc_joins:
  local:
    per_second: 0.2
    burst_count: 15
  remote:
    per_second: 0.03
    burst_count: 12

rc_joins_per_room

This option allows admins to ratelimit joins to a room based on the number of recent joins (local or remote) to that room. It is intended to mitigate mass-join spam waves which target multiple homeservers.

By default, one join is permitted to a room every second, with an accumulating buffer of up to ten instantaneous joins.

Example configuration (default values):

rc_joins_per_room:
  per_second: 1
  burst_count: 10

Added in Synapse 1.64.0.

rc_3pid_validation

This option ratelimits how often a user or IP can attempt to validate a 3PID. Defaults to per_second: 0.003, burst_count: 5.

Example configuration:

rc_3pid_validation:
  per_second: 0.003
  burst_count: 5

rc_invites

This option sets ratelimiting how often invites can be sent in a room or to a specific user. per_room defaults to per_second: 0.3, burst_count: 10 and per_user defaults to per_second: 0.003, burst_count: 5.

Client requests that invite user(s) when creating a room will count against the rc_invites.per_room limit, whereas client requests to invite a single user to a room will count against both the rc_invites.per_user and rc_invites.per_room limits.

Federation requests to invite a user will count against the rc_invites.per_user limit only, as Synapse presumes ratelimiting by room will be done by the sending server.

The rc_invites.per_user limit applies to the receiver of the invite, rather than the sender, meaning that a rc_invite.per_user.burst_count of 5 mandates that a single user cannot receive more than a burst of 5 invites at a time.

In contrast, the rc_invites.per_issuer limit applies to the issuer of the invite, meaning that a rc_invite.per_issuer.burst_count of 5 mandates that single user cannot send more than a burst of 5 invites at a time.

Changed in version 1.63: added the per_issuer limit.

Example configuration:

rc_invites:
  per_room:
    per_second: 0.5
    burst_count: 5
  per_user:
    per_second: 0.004
    burst_count: 3
  per_issuer:
    per_second: 0.5
    burst_count: 5

rc_third_party_invite

This option ratelimits 3PID invites (i.e. invites sent to a third-party ID such as an email address or a phone number) based on the account that's sending the invite. Defaults to per_second: 0.2, burst_count: 10.

Example configuration:

rc_third_party_invite:
  per_second: 0.2
  burst_count: 10

rc_federation

Defines limits on federation requests.

The rc_federation configuration has the following sub-options:

• window_size: window size in milliseconds. Defaults to 1000.
• sleep_limit: number of federation requests from a single server in a window before the server will delay processing the request. Defaults to 10.
• sleep_delay: duration in milliseconds to delay processing events from remote servers by if they go over the sleep limit. Defaults to 500.
• reject_limit: maximum number of concurrent federation requests allowed from a single server. Defaults to 50.
• concurrent: number of federation requests to concurrently process from a single server. Defaults to 3.

Example configuration:

rc_federation:
  window_size: 750
  sleep_limit: 15
  sleep_delay: 400
  reject_limit: 40
  concurrent: 5

federation_rr_transactions_per_room_per_second

Sets outgoing federation transaction frequency for sending read-receipts, per-room.

If we end up trying to send out more read-receipts, they will get buffered up into fewer transactions. Defaults to 50.

Example configuration:

federation_rr_transactions_per_room_per_second: 40

Media Store

Config options related to Synapse's media store.

enable_media_repo

Enable the media store service in the Synapse master. Defaults to true. Set to false if you are using a separate media store worker.

Example configuration:

enable_media_repo: false

media_store_path

Directory where uploaded images and attachments are stored.

Example configuration:

media_store_path: "DATADIR/media_store"

media_storage_providers

Media storage providers allow media to be stored in different locations. Defaults to none. Associated sub-options are:

• module: type of resource, e.g. file_system.
• store_local: whether to store newly uploaded local files
• store_remote: whether to store newly downloaded local files
• store_synchronous: whether to wait for successful storage for local uploads
• config: sets a path to the resource through the directory option

Example configuration:

media_storage_providers:
  - module: file_system
    store_local: false
    store_remote: false
    store_synchronous: false
    config:
       directory: /mnt/some/other/directory

max_upload_size

The largest allowed upload size in bytes.

If you are using a reverse proxy you may also need to set this value in your reverse proxy's config. Defaults to 50M. Notably Nginx has a small max body size by default. See here for more on using a reverse proxy with Synapse.

Example configuration:

max_upload_size: 60M

max_image_pixels

Maximum number of pixels that will be thumbnailed. Defaults to 32M.

Example configuration:

max_image_pixels: 35M

dynamic_thumbnails

Whether to generate new thumbnails on the fly to precisely match the resolution requested by the client. If true then whenever a new resolution is requested by the client the server will generate a new thumbnail. If false the server will pick a thumbnail from a precalculated list. Defaults to false.

Example configuration:

dynamic_thumbnails: true

thumbnail_sizes

List of thumbnails to precalculate when an image is uploaded. Associated sub-options are:

• width
• height
• method: i.e. crop, scale, etc.

Example configuration:

thumbnail_sizes:
  - width: 32
    height: 32
    method: crop
  - width: 96
    height: 96
    method: crop
  - width: 320
    height: 240
    method: scale
  - width: 640
    height: 480
    method: scale
  - width: 800
    height: 600
    method: scale

media_retention

Controls whether local media and entries in the remote media cache (media that is downloaded from other homeservers) should be removed under certain conditions, typically for the purpose of saving space.

Purging media files will be the carried out by the media worker (that is, the worker that has the enable_media_repo homeserver config option set to 'true'). This may be the main process.

The media_retention.local_media_lifetime and media_retention.remote_media_lifetime config options control whether media will be purged if it has not been accessed in a given amount of time. Note that media is 'accessed' when loaded in a room in a client, or otherwise downloaded by a local or remote user. If the media has never been accessed, the media's creation time is used instead. Both thumbnails and the original media will be removed. If either of these options are unset, then media of that type will not be purged.

Local or cached remote media that has been quarantined will not be deleted. Similarly, local media that has been marked as protected from quarantine will not be deleted.

Example configuration:

media_retention:
    local_media_lifetime: 90d
    remote_media_lifetime: 14d

url_preview_enabled

This setting determines whether the preview URL API is enabled. It is disabled by default. Set to true to enable. If enabled you must specify a url_preview_ip_range_blacklist blacklist.

Example configuration:

url_preview_enabled: true

url_preview_ip_range_blacklist

List of IP address CIDR ranges that the URL preview spider is denied from accessing. There are no defaults: you must explicitly specify a list for URL previewing to work. You should specify any internal services in your network that you do not want synapse to try to connect to, otherwise anyone in any Matrix room could cause your synapse to issue arbitrary GET requests to your internal services, causing serious security issues.

(0.0.0.0 and :: are always blacklisted, whether or not they are explicitly listed here, since they correspond to unroutable addresses.)

This must be specified if url_preview_enabled is set. It is recommended that you use the following example list as a starting point.

Note: The value is ignored when an HTTP proxy is in use.

Example configuration:

url_preview_ip_range_blacklist:
  - '127.0.0.0/8'
  - '10.0.0.0/8'
  - '172.16.0.0/12'
  - '192.168.0.0/16'
  - '100.64.0.0/10'
  - '192.0.0.0/24'
  - '169.254.0.0/16'
  - '192.88.99.0/24'
  - '198.18.0.0/15'
  - '192.0.2.0/24'
  - '198.51.100.0/24'
  - '203.0.113.0/24'
  - '224.0.0.0/4'
  - '::1/128'
  - 'fe80::/10'
  - 'fc00::/7'
  - '2001:db8::/32'
  - 'ff00::/8'
  - 'fec0::/10'

url_preview_ip_range_whitelist

This option sets a list of IP address CIDR ranges that the URL preview spider is allowed to access even if they are specified in url_preview_ip_range_blacklist. This is useful for specifying exceptions to wide-ranging blacklisted target IP ranges - e.g. for enabling URL previews for a specific private website only visible in your network. Defaults to none.

Example configuration:

url_preview_ip_range_whitelist:
   - '192.168.1.1'

url_preview_url_blacklist

Optional list of URL matches that the URL preview spider is denied from accessing. You should use url_preview_ip_range_blacklist in preference to this, otherwise someone could define a public DNS entry that points to a private IP address and circumvent the blacklist. This is more useful if you know there is an entire shape of URL that you know that will never want synapse to try to spider.

Each list entry is a dictionary of url component attributes as returned by urlparse.urlsplit as applied to the absolute form of the URL. See here for more information. Some examples are:

• username
• netloc
• scheme
• path

The values of the dictionary are treated as a filename match pattern applied to that component of URLs, unless they start with a ^ in which case they are treated as a regular expression match. If all the specified component matches for a given list item succeed, the URL is blacklisted.

Example configuration:

url_preview_url_blacklist:
  # blacklist any URL with a username in its URI
  - username: '*'

  # blacklist all *.google.com URLs
  - netloc: 'google.com'
  - netloc: '*.google.com'

  # blacklist all plain HTTP URLs
  - scheme: 'http'

  # blacklist http(s)://www.acme.com/foo
  - netloc: 'www.acme.com'
    path: '/foo'

  # blacklist any URL with a literal IPv4 address
  - netloc: '^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$'

max_spider_size

The largest allowed URL preview spidering size in bytes. Defaults to 10M.

Example configuration:

max_spider_size: 8M

url_preview_accept_language

A list of values for the Accept-Language HTTP header used when downloading webpages during URL preview generation. This allows Synapse to specify the preferred languages that URL previews should be in when communicating with remote servers.

Each value is a IETF language tag; a 2-3 letter identifier for a language, optionally followed by subtags separated by '-', specifying a country or region variant.

Multiple values can be provided, and a weight can be added to each by using quality value syntax (;q=). '*' translates to any language.

Defaults to "en".

Example configuration:

 url_preview_accept_language:
   - 'en-UK'
   - 'en-US;q=0.9'
   - 'fr;q=0.8'
   - '*;q=0.7'

oembed

oEmbed allows for easier embedding content from a website. It can be used for generating URLs previews of services which support it. A default list of oEmbed providers is included with Synapse. Set disable_default_providers to true to disable using these default oEmbed URLs. Use additional_providers to specify additional files with oEmbed configuration (each should be in the form of providers.json). By default this list is empty.

Example configuration:

oembed:
  disable_default_providers: true
  additional_providers:
    - oembed/my_providers.json

Captcha

See here for full details on setting up captcha.

recaptcha_public_key

This homeserver's ReCAPTCHA public key. Must be specified if enable_registration_captcha is enabled.

Example configuration:

recaptcha_public_key: "YOUR_PUBLIC_KEY"

recaptcha_private_key

This homeserver's ReCAPTCHA private key. Must be specified if enable_registration_captcha is enabled.

Example configuration:

recaptcha_private_key: "YOUR_PRIVATE_KEY"

enable_registration_captcha

Set to true to require users to complete a CAPTCHA test when registering an account. Requires a valid ReCaptcha public/private key. Defaults to false.

Note that enable_registration must also be set to allow account registration.

Example configuration:

enable_registration_captcha: true

recaptcha_siteverify_api

The API endpoint to use for verifying m.login.recaptcha responses. Defaults to https://www.recaptcha.net/recaptcha/api/siteverify.

Example configuration:

recaptcha_siteverify_api: "https://my.recaptcha.site"

TURN

Options related to adding a TURN server to Synapse.

turn_uris

The public URIs of the TURN server to give to clients.

Example configuration:

turn_uris: [turn:example.org]

turn_shared_secret

The shared secret used to compute passwords for the TURN server.

Example configuration:

turn_shared_secret: "YOUR_SHARED_SECRET"

turn_username and turn_password

The Username and password if the TURN server needs them and does not use a token.

Example configuration:

turn_username: "TURNSERVER_USERNAME"
turn_password: "TURNSERVER_PASSWORD"

turn_user_lifetime

How long generated TURN credentials last. Defaults to 1h.

Example configuration:

turn_user_lifetime: 2h

turn_allow_guests

Whether guests should be allowed to use the TURN server. This defaults to true, otherwise VoIP will be unreliable for guests. However, it does introduce a slight security risk as it allows users to connect to arbitrary endpoints without having first signed up for a valid account (e.g. by passing a CAPTCHA).

Example configuration:

turn_allow_guests: false

Registration

Registration can be rate-limited using the parameters in the Ratelimiting section of this manual.

enable_registration

Enable registration for new users. Defaults to false.

It is highly recommended that if you enable registration, you set one or more or the following options, to avoid abuse of your server by "bots":

• enable_registration_captcha
• registrations_require_3pid
• registration_requires_token

(In order to enable registration without any verification, you must also set enable_registration_without_verification.)

Note that even if this setting is disabled, new accounts can still be created via the admin API if registration_shared_secret is set.

Example configuration:

enable_registration: true

enable_registration_without_verification

Enable registration without email or captcha verification. Note: this option is not recommended, as registration without verification is a known vector for spam and abuse. Defaults to false. Has no effect unless enable_registration is also enabled.

Example configuration:

enable_registration_without_verification: true

registrations_require_3pid

If this is set, users must provide all of the specified types of 3PID when registering an account.

Note that enable_registration must also be set to allow account registration.

Example configuration:

registrations_require_3pid:
  - email
  - msisdn

disable_msisdn_registration

Explicitly disable asking for MSISDNs from the registration flow (overrides registrations_require_3pid if MSISDNs are set as required).

Example configuration:

disable_msisdn_registration: true

allowed_local_3pids

Mandate that users are only allowed to associate certain formats of 3PIDs with accounts on this server, as specified by the medium and pattern sub-options.

Example configuration:

allowed_local_3pids:
  - medium: email
    pattern: '^[^@]+@matrix\.org$'
  - medium: email
    pattern: '^[^@]+@vector\.im$'
  - medium: msisdn
    pattern: '\+44'

enable_3pid_lookup

Enable 3PIDs lookup requests to identity servers from this server. Defaults to true.

Example configuration:

enable_3pid_lookup: false

registration_requires_token

Require users to submit a token during registration. Tokens can be managed using the admin API. Disabling this option will not delete any tokens previously generated. Defaults to false. Set to true to enable.

Note that enable_registration must also be set to allow account registration.

Example configuration:

registration_requires_token: true

registration_shared_secret

If set, allows registration of standard or admin accounts by anyone who has the shared secret, even if enable_registration is not set.

This is primarily intended for use with the register_new_matrix_user script (see Registering a user); however, the interface is documented.

See also registration_shared_secret_path.

Example configuration:

registration_shared_secret: <PRIVATE STRING>

registration_shared_secret_path

An alternative to registration_shared_secret: allows the shared secret to be specified in an external file.

The file should be a plain text file, containing only the shared secret.

If this file does not exist, Synapse will create a new signing key on startup and store it in this file.

Example configuration:

registration_shared_secret_file: /path/to/secrets/file

Added in Synapse 1.67.0.

bcrypt_rounds

Set the number of bcrypt rounds used to generate password hash. Larger numbers increase the work factor needed to generate the hash. The default number is 12 (which equates to 2^12 rounds). N.B. that increasing this will exponentially increase the time required to register or login - e.g. 24 => 2^24 rounds which will take >20 mins. Example configuration:

bcrypt_rounds: 14

allow_guest_access

Allows users to register as guests without a password/email/etc, and participate in rooms hosted on this server which have been made accessible to anonymous users. Defaults to false.

Example configuration:

allow_guest_access: true

default_identity_server

The identity server which we suggest that clients should use when users log in on this server.

(By default, no suggestion is made, so it is left up to the client. This setting is ignored unless public_baseurl is also explicitly set.)

Example configuration:

default_identity_server: https://matrix.org

account_threepid_delegates

Delegate verification of phone numbers to an identity server.

When a user wishes to add a phone number to their account, we need to verify that they actually own that phone number, which requires sending them a text message (SMS). Currently Synapse does not support sending those texts itself and instead delegates the task to an identity server. The base URI for the identity server to be used is specified by the account_threepid_delegates.msisdn option.

If this is left unspecified, Synapse will not allow users to add phone numbers to their account.

(Servers handling the these requests must answer the /requestToken endpoints defined by the Matrix Identity Service API specification.)

Deprecated in Synapse 1.64.0: The email option is deprecated.

Removed in Synapse 1.66.0: The email option has been removed. If present, Synapse will report a configuration error on startup.

Example configuration:

account_threepid_delegates:
    msisdn: http://localhost:8090  # Delegate SMS sending to this local process

enable_set_displayname

Whether users are allowed to change their displayname after it has been initially set. Useful when provisioning users based on the contents of a third-party directory.

Does not apply to server administrators. Defaults to true.

Example configuration:

enable_set_displayname: false

enable_set_avatar_url

Whether users are allowed to change their avatar after it has been initially set. Useful when provisioning users based on the contents of a third-party directory.

Does not apply to server administrators. Defaults to true.

Example configuration:

enable_set_avatar_url: false

enable_3pid_changes

Whether users can change the third-party IDs associated with their accounts (email address and msisdn).

Defaults to true.

Example configuration:

enable_3pid_changes: false

auto_join_rooms

Users who register on this homeserver will automatically be joined to the rooms listed under this option.

By default, any room aliases included in this list will be created as a publicly joinable room when the first user registers for the homeserver. If the room already exists, make certain it is a publicly joinable room, i.e. the join rule of the room must be set to 'public'. You can find more options relating to auto-joining rooms below.

Example configuration:

auto_join_rooms:
  - "#exampleroom:example.com"
  - "#anotherexampleroom:example.com"

autocreate_auto_join_rooms

Where auto_join_rooms are specified, setting this flag ensures that the rooms exist by creating them when the first user on the homeserver registers.

By default the auto-created rooms are publicly joinable from any federated server. Use the autocreate_auto_join_rooms_federated and autocreate_auto_join_room_preset settings to customise this behaviour.

Setting to false means that if the rooms are not manually created, users cannot be auto-joined since they do not exist.

Defaults to true.

Example configuration:

autocreate_auto_join_rooms: false

autocreate_auto_join_rooms_federated

Whether the rooms listen in auto_join_rooms that are auto-created are available via federation. Only has an effect if autocreate_auto_join_rooms is true.

Note that whether a room is federated cannot be modified after creation.

Defaults to true: the room will be joinable from other servers. Set to false to prevent users from other homeservers from joining these rooms.

Example configuration:

autocreate_auto_join_rooms_federated: false

autocreate_auto_join_room_preset

The room preset to use when auto-creating one of auto_join_rooms. Only has an effect if autocreate_auto_join_rooms is true.

Possible values for this option are:

• "public_chat": the room is joinable by anyone, including federated servers if autocreate_auto_join_rooms_federated is true (the default).
• "private_chat": an invitation is required to join these rooms.
• "trusted_private_chat": an invitation is required to join this room and the invitee is assigned a power level of 100 upon joining the room.

If a value of "private_chat" or "trusted_private_chat" is used then auto_join_mxid_localpart must also be configured.

Defaults to "public_chat".

Example configuration:

autocreate_auto_join_room_preset: private_chat

auto_join_mxid_localpart

The local part of the user id which is used to create auto_join_rooms if autocreate_auto_join_rooms is true. If this is not provided then the initial user account that registers will be used to create the rooms.

The user id is also used to invite new users to any auto-join rooms which are set to invite-only.

It must be configured if autocreate_auto_join_room_preset is set to "private_chat" or "trusted_private_chat".

Note that this must be specified in order for new users to be correctly invited to any auto-join rooms which have been set to invite-only (either at the time of creation or subsequently).

Note that, if the room already exists, this user must be joined and have the appropriate permissions to invite new members.

Example configuration:

auto_join_mxid_localpart: system

auto_join_rooms_for_guests

When auto_join_rooms is specified, setting this flag to false prevents guest accounts from being automatically joined to the rooms.

Defaults to true.

Example configuration:

auto_join_rooms_for_guests: false

inhibit_user_in_use_error

Whether to inhibit errors raised when registering a new account if the user ID already exists. If turned on, requests to /register/available will always show a user ID as available, and Synapse won't raise an error when starting a registration with a user ID that already exists. However, Synapse will still raise an error if the registration completes and the username conflicts.

Defaults to false.

Example configuration:

inhibit_user_in_use_error: true

User session management

session_lifetime

Time that a user's session remains valid for, after they log in.

Note that this is not currently compatible with guest logins.

Note also that this is calculated at login time: changes are not applied retrospectively to users who have already logged in.

By default, this is infinite.

Example configuration:

session_lifetime: 24h

refresh_access_token_lifetime

Time that an access token remains valid for, if the session is using refresh tokens.

For more information about refresh tokens, please see the manual.

Note that this only applies to clients which advertise support for refresh tokens.

Note also that this is calculated at login time and refresh time: changes are not applied to existing sessions until they are refreshed.

By default, this is 5 minutes.

Example configuration:

refreshable_access_token_lifetime: 10m

refresh_token_lifetime: 24h

Time that a refresh token remains valid for (provided that it is not exchanged for another one first). This option can be used to automatically log-out inactive sessions. Please see the manual for more information.

Note also that this is calculated at login time and refresh time: changes are not applied to existing sessions until they are refreshed.

By default, this is infinite.

Example configuration:

refresh_token_lifetime: 24h

nonrefreshable_access_token_lifetime

Time that an access token remains valid for, if the session is NOT using refresh tokens.

Please note that not all clients support refresh tokens, so setting this to a short value may be inconvenient for some users who will then be logged out frequently.

Note also that this is calculated at login time: changes are not applied retrospectively to existing sessions for users that have already logged in.

By default, this is infinite.

Example configuration:

nonrefreshable_access_token_lifetime: 24h

Metrics

Config options related to metrics.

enable_metrics

Set to true to enable collection and rendering of performance metrics. Defaults to false.

Example configuration:

enable_metrics: true

sentry

Use this option to enable sentry integration. Provide the DSN assigned to you by sentry with the dsn setting.

NOTE: While attempts are made to ensure that the logs don't contain any sensitive information, this cannot be guaranteed. By enabling this option the sentry server may therefore receive sensitive information, and it in turn may then disseminate sensitive information through insecure notification channels if so configured.

Example configuration:

sentry:
    dsn: "..."

metrics_flags

Flags to enable Prometheus metrics which are not suitable to be enabled by default, either for performance reasons or limited use. Currently the only option is known_servers, which publishes synapse_federation_known_servers, a gauge of the number of servers this homeserver knows about, including itself. May cause performance problems on large homeservers.

Example configuration:

metrics_flags:
    known_servers: true

report_stats

Whether or not to report homeserver usage statistics. This is originally set when generating the config. Set this option to true or false to change the current behavior. See Reporting Homeserver Usage Statistics for information on what data is reported.

Statistics will be reported 5 minutes after Synapse starts, and then every 3 hours after that.

Example configuration:

report_stats: true

report_stats_endpoint

The endpoint to report homeserver usage statistics to. Defaults to https://matrix.org/report-usage-stats/push

Example configuration:

report_stats_endpoint: https://example.com/report-usage-stats/push

API Configuration

Config settings related to the client/server API

room_prejoin_state

Controls for the state that is shared with users who receive an invite to a room. By default, the following state event types are shared with users who receive invites to the room:

• m.room.join_rules
• m.room.canonical_alias
• m.room.avatar
• m.room.encryption
• m.room.name
• m.room.create
• m.room.topic

To change the default behavior, use the following sub-options:

• disable_default_event_types: set to true to disable the above defaults. If this is enabled, only the event types listed in additional_event_types are shared. Defaults to false.
• additional_event_types: Additional state event types to share with users when they are invited to a room. By default, this list is empty (so only the default event types are shared).

Example configuration:

room_prejoin_state:
   disable_default_event_types: true
   additional_event_types:
     - org.example.custom.event.type
     - m.room.join_rules

track_puppeted_user_ips

We record the IP address of clients used to access the API for various reasons, including displaying it to the user in the "Where you're signed in" dialog.

By default, when puppeting another user via the admin API, the client IP address is recorded against the user who created the access token (ie, the admin user), and not the puppeted user.

Set this option to true to also record the IP address against the puppeted user. (This also means that the puppeted user will count as an "active" user for the purpose of monthly active user tracking - see limit_usage_by_mau etc above.)

Example configuration:

track_puppeted_user_ips: true

app_service_config_files

A list of application service config files to use.

Example configuration:

app_service_config_files:
  - app_service_1.yaml
  - app_service_2.yaml

track_appservice_user_ips

Defaults to false. Set to true to enable tracking of application service IP addresses. Implicitly enables MAU tracking for application service users.

Example configuration:

track_appservice_user_ips: true

macaroon_secret_key

A secret which is used to sign

• access token for guest users,
• short-term login token used during SSO logins (OIDC or SAML2) and
• token used for unsubscribing from email notifications.

If none is specified, the registration_shared_secret is used, if one is given; otherwise, a secret key is derived from the signing key.

Example configuration:

macaroon_secret_key: <PRIVATE STRING>

form_secret

A secret which is used to calculate HMACs for form values, to stop falsification of values. Must be specified for the User Consent forms to work.

Example configuration:

form_secret: <PRIVATE STRING>

Signing Keys

Config options relating to signing keys

signing_key_path

Path to the signing key to sign events and federation requests with.

New in Synapse 1.67: If this file does not exist, Synapse will create a new signing key on startup and store it in this file.

Example configuration:

signing_key_path: "CONFDIR/SERVERNAME.signing.key"

old_signing_keys

The keys that the server used to sign messages with but won't use to sign new messages. For each key, key should be the base64-encoded public key, and expired_tsshould be the time (in milliseconds since the unix epoch) that it was last used.

It is possible to build an entry from an old signing.key file using the export_signing_key script which is provided with synapse.

Example configuration:

old_signing_keys:
  "ed25519:id": { key: "base64string", expired_ts: 123456789123 }

key_refresh_interval

How long key response published by this server is valid for. Used to set the valid_until_ts in /key/v2 APIs. Determines how quickly servers will query to check which keys are still valid. Defaults to 1d.

Example configuration:

key_refresh_interval: 2d

trusted_key_servers

The trusted servers to download signing keys from.

When we need to fetch a signing key, each server is tried in parallel.

Normally, the connection to the key server is validated via TLS certificates. Additional security can be provided by configuring a verify key, which will make synapse check that the response is signed by that key.

This setting supercedes an older setting named perspectives. The old format is still supported for backwards-compatibility, but it is deprecated.

trusted_key_servers defaults to matrix.org, but using it will generate a warning on start-up. To suppress this warning, set suppress_key_server_warning to true.

Options for each entry in the list include:

• server_name: the name of the server. Required.
• verify_keys: an optional map from key id to base64-encoded public key. If specified, we will check that the response is signed by at least one of the given keys.
• accept_keys_insecurely: a boolean. Normally, if verify_keys is unset, and federation_verify_certificates is not true, synapse will refuse to start, because this would allow anyone who can spoof DNS responses to masquerade as the trusted key server. If you know what you are doing and are sure that your network environment provides a secure connection to the key server, you can set this to true to override this behaviour.

Example configuration #1:

trusted_key_servers:
  - server_name: "my_trusted_server.example.com"
    verify_keys:
      "ed25519:auto": "abcdefghijklmnopqrstuvwxyzabcdefghijklmopqr"
  - server_name: "my_other_trusted_server.example.com"

Example configuration #2:

trusted_key_servers:
  - server_name: "matrix.org"

suppress_key_server_warning

Set the following to true to disable the warning that is emitted when the trusted_key_servers include 'matrix.org'. See above.

Example configuration:

suppress_key_server_warning: true

key_server_signing_keys_path

The signing keys to use when acting as a trusted key server. If not specified defaults to the server signing key.

Can contain multiple keys, one per line.

Example configuration:

key_server_signing_keys_path: "key_server_signing_keys.key"

Single sign-on integration

The following settings can be used to make Synapse use a single sign-on provider for authentication, instead of its internal password database.

You will probably also want to set the following options to false to disable the regular login/registration flows:

• enable_registration
• password_config.enabled

saml2_config

Enable SAML2 for registration and login. Uses pysaml2. To learn more about pysaml and to find a full list options for configuring pysaml, read the docs here.

At least one of sp_config or config_path must be set in this section to enable SAML login. You can either put your entire pysaml config inline using the sp_config option, or you can specify a path to a psyaml config file with the sub-option config_path. This setting has the following sub-options:

• sp_config: the configuration for the pysaml2 Service Provider. See pysaml2 docs for format of config. Default values will be used for the entityid and service settings, so it is not normally necessary to specify them unless you need to override them. Here are a few useful sub-options for configuring pysaml:
  • metadata: Point this to the IdP's metadata. You must provide either a local file via the local attribute or (preferably) a URL via the remote attribute.
  • accepted_time_diff: 3: Allowed clock difference in seconds between the homeserver and IdP. Defaults to 0.
  • service: By default, the user has to go to our login page first. If you'd like to allow IdP-initiated login, set allow_unsolicited to true under sp in the service section.
• config_path: specify a separate pysaml2 configuration file thusly: config_path: "CONFDIR/sp_conf.py"
• saml_session_lifetime: The lifetime of a SAML session. This defines how long a user has to complete the authentication process, if allow_unsolicited is unset. The default is 15 minutes.
• user_mapping_provider: Using this option, an external module can be provided as a custom solution to mapping attributes returned from a saml provider onto a matrix user. The user_mapping_provider has the following attributes:
  • module: The custom module's class.
  • config: Custom configuration values for the module. Use the values provided in the example if you are using the built-in user_mapping_provider, or provide your own config values for a custom class if you are using one. This section will be passed as a Python dictionary to the module's parse_config method. The built-in provider takes the following two options:
    • mxid_source_attribute: The SAML attribute (after mapping via the attribute maps) to use to derive the Matrix ID from. It is 'uid' by default. Note: This used to be configured by the saml2_config.mxid_source_attribute option. If that is still defined, its value will be used instead.
    • mxid_mapping: The mapping system to use for mapping the saml attribute onto a matrix ID. Options include: hexencode (which maps unpermitted characters to '=xx') and dotreplace (which replaces unpermitted characters with '.'). The default is hexencode. Note: This used to be configured by the saml2_config.mxid_mapping option. If that is still defined, its value will be used instead.
• grandfathered_mxid_source_attribute: In previous versions of synapse, the mapping from SAML attribute to MXID was always calculated dynamically rather than stored in a table. For backwards- compatibility, we will look for user_ids matching such a pattern before creating a new account. This setting controls the SAML attribute which will be used for this backwards-compatibility lookup. Typically it should be 'uid', but if the attribute maps are changed, it may be necessary to change it. The default is 'uid'.
• attribute_requirements: It is possible to configure Synapse to only allow logins if SAML attributes match particular values. The requirements can be listed under attribute_requirements as shown in the example. All of the listed attributes must match for the login to be permitted.
• idp_entityid: If the metadata XML contains multiple IdP entities then the idp_entityid option must be set to the entity to redirect users to. Most deployments only have a single IdP entity and so should omit this option.

Once SAML support is enabled, a metadata file will be exposed at https://<server>:<port>/_synapse/client/saml2/metadata.xml, which you may be able to use to configure your SAML IdP with. Alternatively, you can manually configure the IdP to use an ACS location of https://<server>:<port>/_synapse/client/saml2/authn_response.

Example configuration:

saml2_config:
  sp_config:
    metadata:
      local: ["saml2/idp.xml"]
      remote:
        - url: https://our_idp/metadata.xml
    accepted_time_diff: 3

    service:
      sp:
        allow_unsolicited: true

    # The examples below are just used to generate our metadata xml, and you
    # may well not need them, depending on your setup. Alternatively you
    # may need a whole lot more detail - see the pysaml2 docs!
    description: ["My awesome SP", "en"]
    name: ["Test SP", "en"]

    ui_info:
      display_name:
        - lang: en
          text: "Display Name is the descriptive name of your service."
      description:
        - lang: en
          text: "Description should be a short paragraph explaining the purpose of the service."
      information_url:
        - lang: en
          text: "https://example.com/terms-of-service"
      privacy_statement_url:
        - lang: en
          text: "https://example.com/privacy-policy"
      keywords:
        - lang: en
          text: ["Matrix", "Element"]
      logo:
        - lang: en
          text: "https://example.com/logo.svg"
          width: "200"
          height: "80"

    organization:
      name: Example com
      display_name:
        - ["Example co", "en"]
      url: "http://example.com"

    contact_person:
      - given_name: Bob
        sur_name: "the Sysadmin"
        email_address": ["admin@example.com"]
        contact_type": technical

  saml_session_lifetime: 5m

  user_mapping_provider:
    # Below options are intended for the built-in provider, they should be
    # changed if using a custom module.
    config:
      mxid_source_attribute: displayName
      mxid_mapping: dotreplace

  grandfathered_mxid_source_attribute: upn

  attribute_requirements:
    - attribute: userGroup
      value: "staff"
    - attribute: department
      value: "sales"

  idp_entityid: 'https://our_idp/entityid'

oidc_providers

List of OpenID Connect (OIDC) / OAuth 2.0 identity providers, for registration and login. See here for information on how to configure these options.

For backwards compatibility, it is also possible to configure a single OIDC provider via an oidc_config setting. This is now deprecated and admins are advised to migrate to the oidc_providers format. (When doing that migration, use oidc for the idp_id to ensure that existing users continue to be recognised.)

Options for each entry include:

• idp_id: a unique identifier for this identity provider. Used internally by Synapse; should be a single word such as 'github'. Note that, if this is changed, users authenticating via that provider will no longer be recognised as the same user! (Use "oidc" here if you are migrating from an old oidc_config configuration.)

• idp_name: A user-facing name for this identity provider, which is used to offer the user a choice of login mechanisms.

• idp_icon: An optional icon for this identity provider, which is presented by clients and Synapse's own IdP picker page. If given, must be an MXC URI of the format mxc:///. (An easy way to obtain such an MXC URI is to upload an image to an (unencrypted) room and then copy the "url" from the source of the event.)

• idp_brand: An optional brand for this identity provider, allowing clients to style the login flow according to the identity provider in question. See the spec for possible options here.

• discover: set to false to disable the use of the OIDC discovery mechanism to discover endpoints. Defaults to true.

• issuer: Required. The OIDC issuer. Used to validate tokens and (if discovery is enabled) to discover the provider's endpoints.

• client_id: Required. oauth2 client id to use.

• client_secret: oauth2 client secret to use. May be omitted if client_secret_jwt_key is given, or if client_auth_method is 'none'.

• client_secret_jwt_key: Alternative to client_secret: details of a key used to create a JSON Web Token to be used as an OAuth2 client secret. If given, must be a dictionary with the following properties:

  • key: a pem-encoded signing key. Must be a suitable key for the algorithm specified. Required unless key_file is given.

  • key_file: the path to file containing a pem-encoded signing key file. Required unless key is given.

  • jwt_header: a dictionary giving properties to include in the JWT header. Must include the key alg, giving the algorithm used to sign the JWT, such as "ES256", using the JWA identifiers in RFC7518.

  • jwt_payload: an optional dictionary giving properties to include in the JWT payload. Normally this should include an iss key.

• client_auth_method: auth method to use when exchanging the token. Valid values are client_secret_basic (default), client_secret_post and none.

• scopes: list of scopes to request. This should normally include the "openid" scope. Defaults to ["openid"].

• authorization_endpoint: the oauth2 authorization endpoint. Required if provider discovery is disabled.

• token_endpoint: the oauth2 token endpoint. Required if provider discovery is disabled.

• userinfo_endpoint: the OIDC userinfo endpoint. Required if discovery is disabled and the 'openid' scope is not requested.

• jwks_uri: URI where to fetch the JWKS. Required if discovery is disabled and the 'openid' scope is used.

• skip_verification: set to 'true' to skip metadata verification. Use this if you are connecting to a provider that is not OpenID Connect compliant. Defaults to false. Avoid this in production.

• user_profile_method: Whether to fetch the user profile from the userinfo endpoint, or to rely on the data returned in the id_token from the token_endpoint. Valid values are: auto or userinfo_endpoint. Defaults to auto, which uses the userinfo endpoint if openid is not included in scopes. Set to userinfo_endpoint to always use the userinfo endpoint.

• allow_existing_users: set to true to allow a user logging in via OIDC to match a pre-existing account instead of failing. This could be used if switching from password logins to OIDC. Defaults to false.

• user_mapping_provider: Configuration for how attributes returned from a OIDC provider are mapped onto a matrix user. This setting has the following sub-properties:

  • module: The class name of a custom mapping module. Default is synapse.handlers.oidc.JinjaOidcMappingProvider. See https://matrix-org.github.io/synapse/latest/sso_mapping_providers.html#openid-mapping-providers for information on implementing a custom mapping provider.

  • config: Configuration for the mapping provider module. This section will be passed as a Python dictionary to the user mapping provider module's parse_config method.

    For the default provider, the following settings are available:

    • subject_claim: name of the claim containing a unique identifier for the user. Defaults to 'sub', which OpenID Connect compliant providers should provide.

    • localpart_template: Jinja2 template for the localpart of the MXID. If this is not set, the user will be prompted to choose their own username (see the documentation for the sso_auth_account_details.html template). This template can use the localpart_from_email filter.

    • confirm_localpart: Whether to prompt the user to validate (or change) the generated localpart (see the documentation for the 'sso_auth_account_details.html' template), instead of registering the account right away.

    • display_name_template: Jinja2 template for the display name to set on first login. If unset, no displayname will be set.

    • email_template: Jinja2 template for the email address of the user. If unset, no email address will be added to the account.

    • extra_attributes: a map of Jinja2 templates for extra attributes to send back to the client during login. Note that these are non-standard and clients will ignore them without modifications.

  When rendering, the Jinja2 templates are given a 'user' variable, which is set to the claims returned by the UserInfo Endpoint and/or in the ID Token.

It is possible to configure Synapse to only allow logins if certain attributes match particular values in the OIDC userinfo. The requirements can be listed under attribute_requirements as shown here:

attribute_requirements:
     - attribute: family_name
       value: "Stephensson"
     - attribute: groups
       value: "admin"

All of the listed attributes must match for the login to be permitted. Additional attributes can be added to userinfo by expanding the scopes section of the OIDC config to retrieve additional information from the OIDC provider.

If the OIDC claim is a list, then the attribute must match any value in the list. Otherwise, it must exactly match the value of the claim. Using the example above, the family_name claim MUST be "Stephensson", but the groups claim MUST contain "admin".

Example configuration:

oidc_providers:
  # Generic example
  #
  - idp_id: my_idp
    idp_name: "My OpenID provider"
    idp_icon: "mxc://example.com/mediaid"
    discover: false
    issuer: "https://accounts.example.com/"
    client_id: "provided-by-your-issuer"
    client_secret: "provided-by-your-issuer"
    client_auth_method: client_secret_post
    scopes: ["openid", "profile"]
    authorization_endpoint: "https://accounts.example.com/oauth2/auth"
    token_endpoint: "https://accounts.example.com/oauth2/token"
    userinfo_endpoint: "https://accounts.example.com/userinfo"
    jwks_uri: "https://accounts.example.com/.well-known/jwks.json"
    skip_verification: true
    user_mapping_provider:
      config:
        subject_claim: "id"
        localpart_template: "{{ user.login }}"
        display_name_template: "{{ user.name }}"
        email_template: "{{ user.email }}"
    attribute_requirements:
      - attribute: userGroup
        value: "synapseUsers"

cas_config

Enable Central Authentication Service (CAS) for registration and login. Has the following sub-options:

• enabled: Set this to true to enable authorization against a CAS server. Defaults to false.
• server_url: The URL of the CAS authorization endpoint.
• displayname_attribute: The attribute of the CAS response to use as the display name. If no name is given here, no displayname will be set.
• required_attributes: It is possible to configure Synapse to only allow logins if CAS attributes match particular values. All of the keys given below must exist and the values must match the given value. Alternately if the given value is None then any value is allowed (the attribute just must exist). All of the listed attributes must match for the login to be permitted.

Example configuration:

cas_config:
  enabled: true
  server_url: "https://cas-server.com"
  displayname_attribute: name
  required_attributes:
    userGroup: "staff"
    department: None

sso

Additional settings to use with single-sign on systems such as OpenID Connect, SAML2 and CAS.

Server admins can configure custom templates for pages related to SSO. See here for more information.

Options include:

• client_whitelist: A list of client URLs which are whitelisted so that the user does not have to confirm giving access to their account to the URL. Any client whose URL starts with an entry in the following list will not be subject to an additional confirmation step after the SSO login is completed. WARNING: An entry such as "https://my.client" is insecure, because it will also match "https://my.client.evil.site", exposing your users to phishing attacks from evil.site. To avoid this, include a slash after the hostname: "https://my.client/". The login fallback page (used by clients that don't natively support the required login flows) is whitelisted in addition to any URLs in this list. By default, this list contains only the login fallback page.
• update_profile_information: Use this setting to keep a user's profile fields in sync with information from the identity provider. Currently only syncing the displayname is supported. Fields are checked on every SSO login, and are updated if necessary. Note that enabling this option will override user profile information, regardless of whether users have opted-out of syncing that information when first signing in. Defaults to false.

Example configuration:

sso:
    client_whitelist:
      - https://riot.im/develop
      - https://my.custom.client/
    update_profile_information: true

jwt_config

JSON web token integration. The following settings can be used to make Synapse JSON web tokens for authentication, instead of its internal password database.

Each JSON Web Token needs to contain a "sub" (subject) claim, which is used as the localpart of the mxid.

Additionally, the expiration time ("exp"), not before time ("nbf"), and issued at ("iat") claims are validated if present.

Note that this is a non-standard login type and client support is expected to be non-existent.

See here for more.

Additional sub-options for this setting include:

• enabled: Set to true to enable authorization using JSON web tokens. Defaults to false.
• secret: This is either the private shared secret or the public key used to decode the contents of the JSON web token. Required if enabled is set to true.
• algorithm: The algorithm used to sign (or HMAC) the JSON web token. Supported algorithms are listed here (section JWS). Required if enabled is set to true.
• subject_claim: Name of the claim containing a unique identifier for the user. Optional, defaults to sub.
• issuer: The issuer to validate the "iss" claim against. Optional. If provided the "iss" claim will be required and validated for all JSON web tokens.
• audiences: A list of audiences to validate the "aud" claim against. Optional. If provided the "aud" claim will be required and validated for all JSON web tokens. Note that if the "aud" claim is included in a JSON web token then validation will fail without configuring audiences.

Example configuration:

jwt_config:
    enabled: true
    secret: "provided-by-your-issuer"
    algorithm: "provided-by-your-issuer"
    subject_claim: "name_of_claim"
    issuer: "provided-by-your-issuer"
    audiences:
        - "provided-by-your-issuer"

password_config

Use this setting to enable password-based logins.

This setting has the following sub-options:

• enabled: Defaults to true. Set to false to disable password authentication. Set to only_for_reauth to allow users with existing passwords to use them to log in and reauthenticate, whilst preventing new users from setting passwords.
• localdb_enabled: Set to false to disable authentication against the local password database. This is ignored if enabled is false, and is only useful if you have other password_providers. Defaults to true.
• pepper: Set the value here to a secret random string for extra security. DO NOT CHANGE THIS AFTER INITIAL SETUP!
• policy: Define and enforce a password policy, such as minimum lengths for passwords, etc. Each parameter is optional. This is an implementation of MSC2000. Parameters are as follows:
  • enabled: Defaults to false. Set to true to enable.
  • minimum_length: Minimum accepted length for a password. Defaults to 0.
  • require_digit: Whether a password must contain at least one digit. Defaults to false.
  • require_symbol: Whether a password must contain at least one symbol. A symbol is any character that's not a number or a letter. Defaults to false.
  • require_lowercase: Whether a password must contain at least one lowercase letter. Defaults to false.
  • require_uppercase: Whether a password must contain at least one uppercase letter. Defaults to false.

Example configuration:

password_config:
   enabled: false
   localdb_enabled: false
   pepper: "EVEN_MORE_SECRET"

   policy:
      enabled: true
      minimum_length: 15
      require_digit: true
      require_symbol: true
      require_lowercase: true
      require_uppercase: true

ui_auth

The amount of time to allow a user-interactive authentication session to be active.

This defaults to 0, meaning the user is queried for their credentials before every action, but this can be overridden to allow a single validation to be re-used. This weakens the protections afforded by the user-interactive authentication process, by allowing for multiple (and potentially different) operations to use the same validation session.

This is ignored for potentially "dangerous" operations (including deactivating an account, modifying an account password, and adding a 3PID).

Use the session_timeout sub-option here to change the time allowed for credential validation.

Example configuration:

ui_auth:
    session_timeout: "15s"

email

Configuration for sending emails from Synapse.

Server admins can configure custom templates for email content. See here for more information.

This setting has the following sub-options:

• smtp_host: The hostname of the outgoing SMTP server to use. Defaults to 'localhost'.

• smtp_port: The port on the mail server for outgoing SMTP. Defaults to 465 if force_tls is true, else 25.

  Changed in Synapse 1.64.0: the default port is now aware of force_tls.

• smtp_user and smtp_pass: Username/password for authentication to the SMTP server. By default, no authentication is attempted.

• force_tls: By default, Synapse connects over plain text and then optionally upgrades to TLS via STARTTLS. If this option is set to true, TLS is used from the start (Implicit TLS), and the option require_transport_security is ignored. It is recommended to enable this if supported by your mail server.

  New in Synapse 1.64.0.

• require_transport_security: Set to true to require TLS transport security for SMTP. By default, Synapse will connect over plain text, and will then switch to TLS via STARTTLS if the SMTP server supports it. If this option is set, Synapse will refuse to connect unless the server supports STARTTLS.

• enable_tls: By default, if the server supports TLS, it will be used, and the server must present a certificate that is valid for 'smtp_host'. If this option is set to false, TLS will not be used.

• notif_from: defines the "From" address to use when sending emails. It must be set if email sending is enabled. The placeholder '%(app)s' will be replaced by the application name, which is normally set in app_name, but may be overridden by the Matrix client application. Note that the placeholder must be written '%(app)s', including the trailing 's'.

• app_name: app_name defines the default value for '%(app)s' in notif_from and email subjects. It defaults to 'Matrix'.

• enable_notifs: Set to true to enable sending emails for messages that the user has missed. Disabled by default.

• notif_for_new_users: Set to false to disable automatic subscription to email notifications for new users. Enabled by default.

• client_base_url: Custom URL for client links within the email notifications. By default links will be based on "https://matrix.to". (This setting used to be called riot_base_url; the old name is still supported for backwards-compatibility but is now deprecated.)

• validation_token_lifetime: Configures the time that a validation email will expire after sending. Defaults to 1h.

• invite_client_location: The web client location to direct users to during an invite. This is passed to the identity server as the org.matrix.web_client_location key. Defaults to unset, giving no guidance to the identity server.

• subjects: Subjects to use when sending emails from Synapse. The placeholder '%(app)s' will be replaced with the value of the app_name setting, or by a value dictated by the Matrix client application. In addition, each subject can use the following placeholders: '%(person)s', which will be replaced by the displayname of the user(s) that sent the message(s), e.g. "Alice and Bob", and '%(room)s', which will be replaced by the name of the room the message(s) have been sent to, e.g. "My super room". In addition, emails related to account administration will can use the '%(server_name)s' placeholder, which will be replaced by the value of the server_name setting in your Synapse configuration.

  Here is a list of subjects for notification emails that can be set:

  • message_from_person_in_room: Subject to use to notify about one message from one or more user(s) in a room which has a name. Defaults to "[%(app)s] You have a message on %(app)s from %(person)s in the %(room)s room..."
  • message_from_person: Subject to use to notify about one message from one or more user(s) in a room which doesn't have a name. Defaults to "[%(app)s] You have a message on %(app)s from %(person)s..."
  • messages_from_person: Subject to use to notify about multiple messages from one or more users in a room which doesn't have a name. Defaults to "[%(app)s] You have messages on %(app)s from %(person)s..."
  • messages_in_room: Subject to use to notify about multiple messages in a room which has a name. Defaults to "[%(app)s] You have messages on %(app)s in the %(room)s room..."
  • messages_in_room_and_others: Subject to use to notify about multiple messages in multiple rooms. Defaults to "[%(app)s] You have messages on %(app)s in the %(room)s room and others..."
  • messages_from_person_and_others: Subject to use to notify about multiple messages from multiple persons in multiple rooms. This is similar to the setting above except it's used when the room in which the notification was triggered has no name. Defaults to "[%(app)s] You have messages on %(app)s from %(person)s and others..."
  • invite_from_person_to_room: Subject to use to notify about an invite to a room which has a name. Defaults to "[%(app)s] %(person)s has invited you to join the %(room)s room on %(app)s..."
  • invite_from_person: Subject to use to notify about an invite to a room which doesn't have a name. Defaults to "[%(app)s] %(person)s has invited you to chat on %(app)s..."
  • password_reset: Subject to use when sending a password reset email. Defaults to "[%(server_name)s] Password reset"
  • email_validation: Subject to use when sending a verification email to assert an address's ownership. Defaults to "[%(server_name)s] Validate your email"

Example configuration:

email:
  smtp_host: mail.server
  smtp_port: 587
  smtp_user: "exampleusername"
  smtp_pass: "examplepassword"
  force_tls: true
  require_transport_security: true
  enable_tls: false
  notif_from: "Your Friendly %(app)s homeserver <noreply@example.com>"
  app_name: my_branded_matrix_server
  enable_notifs: true
  notif_for_new_users: false
  client_base_url: "http://localhost/riot"
  validation_token_lifetime: 15m
  invite_client_location: https://app.element.io

  subjects:
    message_from_person_in_room: "[%(app)s] You have a message on %(app)s from %(person)s in the %(room)s room..."
    message_from_person: "[%(app)s] You have a message on %(app)s from %(person)s..."
    messages_from_person: "[%(app)s] You have messages on %(app)s from %(person)s..."
    messages_in_room: "[%(app)s] You have messages on %(app)s in the %(room)s room..."
    messages_in_room_and_others: "[%(app)s] You have messages on %(app)s in the %(room)s room and others..."
    messages_from_person_and_others: "[%(app)s] You have messages on %(app)s from %(person)s and others..."
    invite_from_person_to_room: "[%(app)s] %(person)s has invited you to join the %(room)s room on %(app)s..."
    invite_from_person: "[%(app)s] %(person)s has invited you to chat on %(app)s..."
    password_reset: "[%(server_name)s] Password reset"
    email_validation: "[%(server_name)s] Validate your email"

Push

Configuration settings related to push notifications

push

This setting defines options for push notifications.

This option has a number of sub-options. They are as follows:

• include_content: Clients requesting push notifications can either have the body of the message sent in the notification poke along with other details like the sender, or just the event ID and room ID (event_id_only). If clients choose the to have the body sent, this option controls whether the notification request includes the content of the event (other details like the sender are still included). If event_id_only is enabled, it has no effect. For modern android devices the notification content will still appear because it is loaded by the app. iPhone, however will send a notification saying only that a message arrived and who it came from. Defaults to true. Set to false to only include the event ID and room ID in push notification payloads.
• group_unread_count_by_room: false: When a push notification is received, an unread count is also sent. This number can either be calculated as the number of unread messages for the user, or the number of rooms the user has unread messages in. Defaults to true, meaning push clients will see the number of rooms with unread messages in them. Set to false to instead send the number of unread messages.

Example configuration:

push:
  include_content: false
  group_unread_count_by_room: false

Rooms

Config options relating to rooms.

encryption_enabled_by_default

Controls whether locally-created rooms should be end-to-end encrypted by default.

Possible options are "all", "invite", and "off". They are defined as:

• "all": any locally-created room
• "invite": any room created with the private_chat or trusted_private_chat room creation presets
• "off": this option will take no effect

The default value is "off".

Note that this option will only affect rooms created after it is set. It will also not affect rooms created by other servers.

Example configuration:

encryption_enabled_by_default_for_room_type: invite

user_directory

This setting defines options related to the user directory.

This option has the following sub-options:

• enabled: Defines whether users can search the user directory. If false then empty responses are returned to all queries. Defaults to true.
• search_all_users: Defines whether to search all users visible to your HS when searching the user directory. If false, search results will only contain users visible in public rooms and users sharing a room with the requester. Defaults to false. NB. If you set this to true, and the last time the user_directory search indexes were (re)built was before Synapse 1.44, you'll have to rebuild the indexes in order to search through all known users. These indexes are built the first time Synapse starts; admins can manually trigger a rebuild via API following the instructions at https://matrix-org.github.io/synapse/latest/usage/administration/admin_api/background_updates.html#run Set to true to return search results containing all known users, even if that user does not share a room with the requester.
• prefer_local_users: Defines whether to prefer local users in search query results. If set to true, local users are more likely to appear above remote users when searching the user directory. Defaults to false.

Example configuration:

user_directory:
    enabled: false
    search_all_users: true
    prefer_local_users: true

user_consent

For detailed instructions on user consent configuration, see here.

Parts of this section are required if enabling the consent resource under listeners, in particular template_dir and version.

• template_dir: gives the location of the templates for the HTML forms. This directory should contain one subdirectory per language (eg, en, fr), and each language directory should contain the policy document (named as .html) and a success page (success.html).

• version: specifies the 'current' version of the policy document. It defines the version to be served by the consent resource if there is no 'v' parameter.

• server_notice_content: if enabled, will send a user a "Server Notice" asking them to consent to the privacy policy. The server_notices section must also be configured for this to work. Notices will not be sent to guest users unless send_server_notice_to_guests is set to true.

• block_events_error, if set, will block any attempts to send events until the user consents to the privacy policy. The value of the setting is used as the text of the error.

• require_at_registration, if enabled, will add a step to the registration process, similar to how captcha works. Users will be required to accept the policy before their account is created.

• policy_name is the display name of the policy users will see when registering for an account. Has no effect unless require_at_registration is enabled. Defaults to "Privacy Policy".

Example configuration:

user_consent:
  template_dir: res/templates/privacy
  version: 1.0
  server_notice_content:
    msgtype: m.text
    body: >-
      To continue using this homeserver you must review and agree to the
      terms and conditions at %(consent_uri)s
  send_server_notice_to_guests: true
  block_events_error: >-
    To continue using this homeserver you must review and agree to the
    terms and conditions at %(consent_uri)s
  require_at_registration: false
  policy_name: Privacy Policy

stats

Settings for local room and user statistics collection. See here for more.

• enabled: Set to false to disable room and user statistics. Note that doing so may cause certain features (such as the room directory) not to work correctly. Defaults to true.

Example configuration:

stats:
  enabled: false

server_notices

Use this setting to enable a room which can be used to send notices from the server to users. It is a special room which users cannot leave; notices in the room come from a special "notices" user id.

If you use this setting, you must define the system_mxid_localpart sub-setting, which defines the id of the user which will be used to send the notices.

Sub-options for this setting include:

• system_mxid_display_name: set the display name of the "notices" user
• system_mxid_avatar_url: set the avatar for the "notices" user
• room_name: set the room name of the server notices room

Example configuration:

server_notices:
  system_mxid_localpart: notices
  system_mxid_display_name: "Server Notices"
  system_mxid_avatar_url: "mxc://server.com/oumMVlgDnLYFaPVkExemNVVZ"
  room_name: "Server Notices"

enable_room_list_search

Set to false to disable searching the public room list. When disabled blocks searching local and remote room lists for local and remote users by always returning an empty list for all queries. Defaults to true.

Example configuration:

enable_room_list_search: false

alias_creation

The alias_creation option controls who is allowed to create aliases on this server.

The format of this option is a list of rules that contain globs that match against user_id, room_id and the new alias (fully qualified with server name). The action in the first rule that matches is taken, which can currently either be "allow" or "deny".

Missing user_id/room_id/alias fields default to "*".

If no rules match the request is denied. An empty list means no one can create aliases.

Options for the rules include:

• user_id: Matches against the creator of the alias. Defaults to "*".
• alias: Matches against the alias being created. Defaults to "*".
• room_id: Matches against the room ID the alias is being pointed at. Defaults to "*"
• action: Whether to "allow" or "deny" the request if the rule matches. Defaults to allow.

Example configuration:

alias_creation_rules:
  - user_id: "bad_user"
    alias: "spammy_alias"
    room_id: "*"
    action: deny

room_list_publication_rules

The room_list_publication_rules option controls who can publish and which rooms can be published in the public room list.

The format of this option is the same as that for alias_creation_rules.

If the room has one or more aliases associated with it, only one of the aliases needs to match the alias rule. If there are no aliases then only rules with alias: * match.

If no rules match the request is denied. An empty list means no one can publish rooms.

Options for the rules include:

• user_id: Matches against the creator of the alias. Defaults to "*".
• alias: Matches against any current local or canonical aliases associated with the room. Defaults to "*".
• room_id: Matches against the room ID being published. Defaults to "*".
• action: Whether to "allow" or "deny" the request if the rule matches. Defaults to allow.

Example configuration:

room_list_publication_rules:
  - user_id: "*"
    alias: "*"
    room_id: "*"
    action: allow

default_power_level_content_override

The default_power_level_content_override option controls the default power levels for rooms.

Useful if you know that your users need special permissions in rooms that they create (e.g. to send particular types of state events without needing an elevated power level). This takes the same shape as the power_level_content_override parameter in the /createRoom API, but is applied before that parameter.

Note that each key provided inside a preset (for example events in the example below) will overwrite all existing defaults inside that key. So in the example below, newly-created private_chat rooms will have no rules for any event types except com.example.foo.

Example configuration:

default_power_level_content_override:
   private_chat: { "events": { "com.example.foo" : 0 } }
   trusted_private_chat: null
   public_chat: null

Opentracing

Configuration options related to Opentracing support.

opentracing

These settings enable and configure opentracing, which implements distributed tracing. This allows you to observe the causal chains of events across servers including requests, key lookups etc., across any server running synapse or any other services which support opentracing (specifically those implemented with Jaeger).

Sub-options include:

• enabled: whether tracing is enabled. Set to true to enable. Disabled by default.
• homeserver_whitelist: The list of homeservers we wish to send and receive span contexts and span baggage. See here for more. This is a list of regexes which are matched against the server_name of the homeserver. By default, it is empty, so no servers are matched.
• force_tracing_for_users: # A list of the matrix IDs of users whose requests will always be traced, even if the tracing system would otherwise drop the traces due to probabilistic sampling. By default, the list is empty.
• jaeger_config: Jaeger can be configured to sample traces at different rates. All configuration options provided by Jaeger can be set here. Jaeger's configuration is mostly related to trace sampling which is documented here.

Example configuration:

opentracing:
    enabled: true
    homeserver_whitelist:
      - ".*"
    force_tracing_for_users:
      - "@user1:server_name"
      - "@user2:server_name"

    jaeger_config:
      sampler:
        type: const
        param: 1
      logging:
        false

Workers

Configuration options related to workers.

send_federation

Controls sending of outbound federation transactions on the main process. Set to false if using a federation sender worker. Defaults to true.

Example configuration:

send_federation: false

federation_sender_instances

It is possible to run multiple federation sender workers, in which case the work is balanced across them. Use this setting to list the senders.

This configuration setting must be shared between all federation sender workers, and if changed all federation sender workers must be stopped at the same time and then started, to ensure that all instances are running with the same config (otherwise events may be dropped).

Example configuration:

federation_sender_instances:
  - federation_sender1

instance_map

When using workers this should be a map from worker name to the HTTP replication listener of the worker, if configured.

Example configuration:

instance_map:
  worker1:
    host: localhost
    port: 8034

stream_writers

Experimental: When using workers you can define which workers should handle event persistence and typing notifications. Any worker specified here must also be in the instance_map.

Example configuration:

stream_writers:
  events: worker1
  typing: worker1

run_background_tasks_on

The worker that is used to run background tasks (e.g. cleaning up expired data). If not provided this defaults to the main process.

Example configuration:

run_background_tasks_on: worker1

worker_replication_secret

A shared secret used by the replication APIs to authenticate HTTP requests from workers.

By default this is unused and traffic is not authenticated.

Example configuration:

worker_replication_secret: "secret_secret"

redis

Configuration for Redis when using workers. This must be enabled when using workers (unless using old style direct TCP configuration). This setting has the following sub-options:

• enabled: whether to use Redis support. Defaults to false.
• host and port: Optional host and port to use to connect to redis. Defaults to localhost and 6379
• password: Optional password if configured on the Redis instance.

Example configuration:

redis:
  enabled: true
  host: localhost
  port: 6379
  password: <secret_password>

Background Updates

Configuration settings related to background updates.

background_updates

Background updates are database updates that are run in the background in batches. The duration, minimum batch size, default batch size, whether to sleep between batches and if so, how long to sleep can all be configured. This is helpful to speed up or slow down the updates. This setting has the following sub-options:

• background_update_duration_ms: How long in milliseconds to run a batch of background updates for. Defaults to 100. Set a different time to change the default.
• sleep_enabled: Whether to sleep between updates. Defaults to true. Set to false to change the default.
• sleep_duration_ms: If sleeping between updates, how long in milliseconds to sleep for. Defaults to 1000. Set a duration to change the default.
• min_batch_size: Minimum size a batch of background updates can be. Must be greater than 0. Defaults to 1. Set a size to change the default.
• default_batch_size: The batch size to use for the first iteration of a new background update. The default is 100. Set a size to change the default.

Example configuration:

background_updates:
    background_update_duration_ms: 500
    sleep_enabled: false
    sleep_duration_ms: 300
    min_batch_size: 10
    default_batch_size: 50

Homeserver Sample Configuration File

Below is a sample homeserver configuration file. The homeserver configuration file can be tweaked to change the behaviour of your homeserver. A restart of the server is generally required to apply any changes made to this file.

Note that the contents below are not intended to be copied and used as the basis for a real homeserver.yaml. Instead, if you are starting from scratch, please generate a fresh config using Synapse by following the instructions in Installation.

Documentation for all configuration options can be found in the Configuration Manual.

# This file is maintained as an up-to-date snapshot of the default
# homeserver.yaml configuration generated by Synapse. You can find a
# complete accounting of possible configuration options at
# https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html
#
# It is *not* intended to be copied and used as the basis for a real
# homeserver.yaml. Instead, if you are starting from scratch, please generate
# a fresh config using Synapse by following the instructions in
# https://matrix-org.github.io/synapse/latest/setup/installation.html.
#
################################################################################

# Configuration file for Synapse.
#
# This is a YAML file: see [1] for a quick introduction. Note in particular
# that *indentation is important*: all the elements of a list or dictionary
# should have the same indentation.
#
# [1] https://docs.ansible.com/ansible/latest/reference_appendices/YAMLSyntax.html
#
# For more information on how to configure Synapse, including a complete accounting of
# each option, go to docs/usage/configuration/config_documentation.md or
# https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html
server_name: "SERVERNAME"
pid_file: DATADIR/homeserver.pid
listeners:
  - port: 8008
    tls: false
    type: http
    x_forwarded: true
    bind_addresses: ['::1', '127.0.0.1']
    resources:
      - names: [client, federation]
        compress: false
database:
  name: sqlite3
  args:
    database: DATADIR/homeserver.db
log_config: "CONFDIR/SERVERNAME.log.config"
media_store_path: DATADIR/media_store
signing_key_path: "CONFDIR/SERVERNAME.signing.key"
trusted_key_servers:
  - server_name: "matrix.org"

Logging Sample Configuration File

Below is a sample logging configuration file. This file can be tweaked to control how your homeserver will output logs. A restart of the server is generally required to apply any changes made to this file. The value of the log_config option in your homeserver config should be the path to this file.

Note that a default logging configuration (shown below) is created automatically alongside the homeserver config when following the installation instructions. It should be named <SERVERNAME>.log.config by default.

# Log configuration for Synapse.
#
# This is a YAML file containing a standard Python logging configuration
# dictionary. See [1] for details on the valid settings.
#
# Synapse also supports structured logging for machine readable logs which can
# be ingested by ELK stacks. See [2] for details.
#
# [1]: https://docs.python.org/3.7/library/logging.config.html#configuration-dictionary-schema
# [2]: https://matrix-org.github.io/synapse/latest/structured_logging.html

version: 1

formatters:
    precise:
        format: '%(asctime)s - %(name)s - %(lineno)d - %(levelname)s - %(request)s - %(message)s'

handlers:
    file:
        class: logging.handlers.TimedRotatingFileHandler
        formatter: precise
        filename: /var/log/matrix-synapse/homeserver.log
        when: midnight
        backupCount: 3  # Does not include the current log file.
        encoding: utf8

    # Default to buffering writes to log file for efficiency.
    # WARNING/ERROR logs will still be flushed immediately, but there will be a
    # delay (of up to `period` seconds, or until the buffer is full with
    # `capacity` messages) before INFO/DEBUG logs get written.
    buffer:
        class: synapse.logging.handlers.PeriodicallyFlushingMemoryHandler
        target: file

        # The capacity is the maximum number of log lines that are buffered
        # before being written to disk. Increasing this will lead to better
        # performance, at the expensive of it taking longer for log lines to
        # be written to disk.
        # This parameter is required.
        capacity: 10

        # Logs with a level at or above the flush level will cause the buffer to
        # be flushed immediately.
        # Default value: 40 (ERROR)
        # Other values: 50 (CRITICAL), 30 (WARNING), 20 (INFO), 10 (DEBUG)
        flushLevel: 30  # Flush immediately for WARNING logs and higher

        # The period of time, in seconds, between forced flushes.
        # Messages will not be delayed for longer than this time.
        # Default value: 5 seconds
        period: 5

    # A handler that writes logs to stderr. Unused by default, but can be used
    # instead of "buffer" and "file" in the logger handlers.
    console:
        class: logging.StreamHandler
        formatter: precise

loggers:
    synapse.storage.SQL:
        # beware: increasing this to DEBUG will make synapse log sensitive
        # information such as access tokens.
        level: INFO

root:
    level: INFO

    # Write logs to the `buffer` handler, which will buffer them together in memory,
    # then write them to a file.
    #
    # Replace "buffer" with "console" to log to stderr instead. (Note that you'll
    # also need to update the configuration for the `twisted` logger above, in
    # this case.)
    #
    handlers: [buffer]

disable_existing_loggers: false

Structured Logging

A structured logging system can be useful when your logs are destined for a machine to parse and process. By maintaining its machine-readable characteristics, it enables more efficient searching and aggregations when consumed by software such as the "ELK stack".

Synapse's structured logging system is configured via the file that Synapse's log_config config option points to. The file should include a formatter which uses the synapse.logging.TerseJsonFormatter class included with Synapse and a handler which uses the above formatter.

There is also a synapse.logging.JsonFormatter option which does not include a timestamp in the resulting JSON. This is useful if the log ingester adds its own timestamp.

A structured logging configuration looks similar to the following:

version: 1

formatters:
    structured:
        class: synapse.logging.TerseJsonFormatter

handlers:
    file:
        class: logging.handlers.TimedRotatingFileHandler
        formatter: structured
        filename: /path/to/my/logs/homeserver.log
        when: midnight
        backupCount: 3  # Does not include the current log file.
        encoding: utf8

loggers:
    synapse:
        level: INFO
        handlers: [remote]
    synapse.storage.SQL:
        level: WARNING

The above logging config will set Synapse as 'INFO' logging level by default, with the SQL layer at 'WARNING', and will log to a file, stored as JSON.

It is also possible to configure Synapse to log to a remote endpoint by using the synapse.logging.RemoteHandler class included with Synapse. It takes the following arguments:

• host: Hostname or IP address of the log aggregator.
• port: Numerical port to contact on the host.
• maximum_buffer: (Optional, defaults to 1000) The maximum buffer size to allow.

A remote structured logging configuration looks similar to the following:

version: 1

formatters:
    structured:
        class: synapse.logging.TerseJsonFormatter

handlers:
    remote:
        class: synapse.logging.RemoteHandler
        formatter: structured
        host: 10.1.2.3
        port: 9999

loggers:
    synapse:
        level: INFO
        handlers: [remote]
    synapse.storage.SQL:
        level: WARNING

The above logging config will set Synapse as 'INFO' logging level by default, with the SQL layer at 'WARNING', and will log JSON formatted messages to a remote endpoint at 10.1.2.3:9999.

Templates

Synapse uses parametrised templates to generate the content of emails it sends and webpages it shows to users.

By default, Synapse will use the templates listed here. Server admins can configure an additional directory for Synapse to look for templates in, allowing them to specify custom templates:

templates:
  custom_template_directory: /path/to/custom/templates/

If this setting is not set, or the files named below are not found within the directory, default templates from within the Synapse package will be used.

Templates that are given variables when being rendered are rendered using Jinja 2. Templates rendered by Jinja 2 can also access two functions on top of the functions already available as part of Jinja 2:

format_ts(value: int, format: str) -> str

Formats a timestamp in milliseconds.

Example: reason.last_sent_ts|format_ts("%c")

mxc_to_http(value: str, width: int, height: int, resize_method: str = "crop") -> str

Turns a mxc:// URL for media content into an HTTP(S) one using the homeserver's public_baseurl configuration setting as the URL's base.

Example: message.sender_avatar_url|mxc_to_http(32,32)

localpart_from_email(address: str) -> str

Returns the local part of an email address (e.g. alice in alice@example.com).

Example: user.email_address|localpart_from_email

Email templates

Below are the templates Synapse will look for when generating the content of an email:

• notif_mail.html and notif_mail.txt: The contents of email notifications of missed events. When rendering, this template is given the following variables:
  • user_display_name: the display name for the user receiving the notification
  • unsubscribe_link: the link users can click to unsubscribe from email notifications
  • summary_text: a summary of the notification(s). The text used can be customised by configuring the various settings in the email.subjects section of the configuration file.
  • rooms: a list of rooms containing events to include in the email. Each element is an object with the following attributes:
    • title: a human-readable name for the room
    • hash: a hash of the ID of the room
    • invite: a boolean, which is True if the room is an invite the user hasn't accepted yet, False otherwise
    • notifs: a list of events, or an empty list if invite is True. Each element is an object with the following attributes:
      • link: a matrix.to link to the event
      • ts: the time in milliseconds at which the event was received
      • messages: a list of messages containing one message before the event, the message in the event, and one message after the event. Each element is an object with the following attributes:
        • event_type: the type of the event
        • is_historical: a boolean, which is False if the message is the one that triggered the notification, True otherwise
        • id: the ID of the event
        • ts: the time in milliseconds at which the event was sent
        • sender_name: the display name for the event's sender
        • sender_avatar_url: the avatar URL (as a mxc:// URL) for the event's sender
        • sender_hash: a hash of the user ID of the sender
        • msgtype: the type of the message
        • body_text_html: html representation of the message
        • body_text_plain: plaintext representation of the message
        • image_url: mxc url of an image, when "msgtype" is "m.image"
    • link: a matrix.to link to the room
    • avator_url: url to the room's avator
  • reason: information on the event that triggered the email to be sent. It's an object with the following attributes:
    • room_id: the ID of the room the event was sent in
    • room_name: a human-readable name for the room the event was sent in
    • now: the current time in milliseconds
    • received_at: the time in milliseconds at which the event was received
    • delay_before_mail_ms: the amount of time in milliseconds Synapse always waits before ever emailing about a notification (to give the user a chance to respond to other push or notice the window)
    • last_sent_ts: the time in milliseconds at which a notification was last sent for an event in this room
    • throttle_ms: the minimum amount of time in milliseconds between two notifications can be sent for this room
• password_reset.html and password_reset.txt: The contents of password reset emails sent by the homeserver. When rendering, these templates are given a link variable which contains the link the user must click in order to reset their password.
• registration.html and registration.txt: The contents of address verification emails sent during registration. When rendering, these templates are given a link variable which contains the link the user must click in order to validate their email address.
• add_threepid.html and add_threepid.txt: The contents of address verification emails sent when an address is added to a Matrix account. When rendering, these templates are given a link variable which contains the link the user must click in order to validate their email address.

HTML page templates for registration and password reset

Below are the templates Synapse will look for when generating pages related to registration and password reset:

• password_reset_confirmation.html: An HTML page that a user will see when they follow the link in the password reset email. The user will be asked to confirm the action before their password is reset. When rendering, this template is given the following variables:
  • sid: the session ID for the password reset
  • token: the token for the password reset
  • client_secret: the client secret for the password reset
• password_reset_success.html and password_reset_failure.html: HTML pages for success and failure that a user will see when they confirm the password reset flow using the page above. When rendering, password_reset_success.html is given no variable, and password_reset_failure.html is given a failure_reason, which contains the reason for the password reset failure.
• registration_success.html and registration_failure.html: HTML pages for success and failure that a user will see when they follow the link in an address verification email sent during registration. When rendering, registration_success.html is given no variable, and registration_failure.html is given a failure_reason, which contains the reason for the registration failure.
• add_threepid_success.html and add_threepid_failure.html: HTML pages for success and failure that a user will see when they follow the link in an address verification email sent when an address is added to a Matrix account. When rendering, add_threepid_success.html is given no variable, and add_threepid_failure.html is given a failure_reason, which contains the reason for the registration failure.

HTML page templates for Single Sign-On (SSO)

Below are the templates Synapse will look for when generating pages related to SSO:

• sso_login_idp_picker.html: HTML page to prompt the user to choose an Identity Provider during login. This is only used if multiple SSO Identity Providers are configured. When rendering, this template is given the following variables:
  • redirect_url: the URL that the user will be redirected to after login.
  • server_name: the homeserver's name.
  • providers: a list of available Identity Providers. Each element is an object with the following attributes:
    • idp_id: unique identifier for the IdP
    • idp_name: user-facing name for the IdP
    • idp_icon: if specified in the IdP config, an MXC URI for an icon for the IdP
    • idp_brand: if specified in the IdP config, a textual identifier for the brand of the IdP The rendered HTML page should contain a form which submits its results back as a GET request, with the following query parameters:
  • redirectUrl: the client redirect URI (ie, the redirect_url passed to the template)
  • idp: the 'idp_id' of the chosen IDP.
• sso_auth_account_details.html: HTML page to prompt new users to enter a userid and confirm other details. This is only shown if the SSO implementation (with any user_mapping_provider) does not return a localpart. When rendering, this template is given the following variables:
  • server_name: the homeserver's name.
  • idp: details of the SSO Identity Provider that the user logged in with: an object with the following attributes:
    • idp_id: unique identifier for the IdP
    • idp_name: user-facing name for the IdP
    • idp_icon: if specified in the IdP config, an MXC URI for an icon for the IdP
    • idp_brand: if specified in the IdP config, a textual identifier for the brand of the IdP
  • user_attributes: an object containing details about the user that we received from the IdP. May have the following attributes:
    • display_name: the user's display name
    • emails: a list of email addresses
    • localpart: the local part of the Matrix user ID to register, if localpart_template is set in the mapping provider configuration (empty string if not) The template should render a form which submits the following fields:
  • username: the localpart of the user's chosen user id
• sso_new_user_consent.html: HTML page allowing the user to consent to the server's terms and conditions. This is only shown for new users, and only if user_consent.require_at_registration is set. When rendering, this template is given the following variables:
  • server_name: the homeserver's name.
  • user_id: the user's matrix proposed ID.
  • user_profile.display_name: the user's proposed display name, if any.
  • consent_version: the version of the terms that the user will be shown
  • terms_url: a link to the page showing the terms. The template should render a form which submits the following fields:
  • accepted_version: the version of the terms accepted by the user (ie, 'consent_version' from the input variables).
• sso_redirect_confirm.html: HTML page for a confirmation step before redirecting back to the client with the login token. When rendering, this template is given the following variables:
  • redirect_url: the URL the user is about to be redirected to.
  • display_url: the same as redirect_url, but with the query parameters stripped. The intention is to have a human-readable URL to show to users, not to use it as the final address to redirect to.
  • server_name: the homeserver's name.
  • new_user: a boolean indicating whether this is the user's first time logging in.
  • user_id: the user's matrix ID.
  • user_profile.avatar_url: an MXC URI for the user's avatar, if any. None if the user has not set an avatar.
  • user_profile.display_name: the user's display name. None if the user has not set a display name.
• sso_auth_confirm.html: HTML page which notifies the user that they are authenticating to confirm an operation on their account during the user interactive authentication process. When rendering, this template is given the following variables:
  • redirect_url: the URL the user is about to be redirected to.
  • description: the operation which the user is being asked to confirm
  • idp: details of the Identity Provider that we will use to confirm the user's identity: an object with the following attributes:
    • idp_id: unique identifier for the IdP
    • idp_name: user-facing name for the IdP
    • idp_icon: if specified in the IdP config, an MXC URI for an icon for the IdP
    • idp_brand: if specified in the IdP config, a textual identifier for the brand of the IdP
• sso_auth_success.html: HTML page shown after a successful user interactive authentication session. Note that this page must include the JavaScript which notifies of a successful authentication (see https://matrix.org/docs/spec/client_server/r0.6.0#fallback). This template has no additional variables.
• sso_auth_bad_user.html: HTML page shown after a user-interactive authentication session which does not map correctly onto the expected user. When rendering, this template is given the following variables:
  • server_name: the homeserver's name.
  • user_id_to_verify: the MXID of the user that we are trying to validate.
• sso_account_deactivated.html: HTML page shown during single sign-on if a deactivated user (according to Synapse's database) attempts to login. This template has no additional variables.
• sso_error.html: HTML page to display to users if something goes wrong during the OpenID Connect authentication process. When rendering, this template is given two variables:
  • error: the technical name of the error
  • error_description: a human-readable message for the error

User Authentication

Synapse supports multiple methods of authenticating users, either out-of-the-box or through custom pluggable authentication modules.

Included in Synapse is support for authenticating users via:

• A username and password.
• An email address and password.
• Single Sign-On through the SAML, Open ID Connect or CAS protocols.
• JSON Web Tokens.
• An administrator's shared secret.

Synapse can additionally be extended to support custom authentication schemes through optional "password auth provider" modules.

Single Sign-On

Synapse supports single sign-on through the SAML, Open ID Connect or CAS protocols. LDAP and other login methods are supported through first and third-party password auth provider modules.

Configuring Synapse to authenticate against an OpenID Connect provider

Synapse can be configured to use an OpenID Connect Provider (OP) for authentication, instead of its own local password database.

Any OP should work with Synapse, as long as it supports the authorization code flow. There are a few options for that:

• start a local OP. Synapse has been tested with Hydra and Dex. Note that for an OP to work, it should be served under a secure (HTTPS) origin. A certificate signed with a self-signed, locally trusted CA should work. In that case, start Synapse with a SSL_CERT_FILE environment variable set to the path of the CA.

• set up a SaaS OP, like Google, Auth0 or Okta. Synapse has been tested with Auth0 and Google.

It may also be possible to use other OAuth2 providers which provide the authorization code grant type, such as Github.

Preparing Synapse

The OpenID integration in Synapse uses the authlib library, which must be installed as follows:

• The relevant libraries are included in the Docker images and Debian packages provided by matrix.org so no further action is needed.

• If you installed Synapse into a virtualenv, run /path/to/env/bin/pip install matrix-synapse[oidc] to install the necessary dependencies.

• For other installation mechanisms, see the documentation provided by the maintainer.

To enable the OpenID integration, you should then add a section to the oidc_providers setting in your configuration file. See the configuration manual for some sample settings, as well as the text below for example configurations for specific providers.

Sample configs

Here are a few configs for providers that should work with Synapse.

Microsoft Azure Active Directory

Azure AD can act as an OpenID Connect Provider. Register a new application under App registrations in the Azure AD management console. The RedirectURI for your application should point to your matrix server: [synapse public baseurl]/_synapse/client/oidc/callback

Go to Certificates & secrets and register a new client secret. Make note of your Directory (tenant) ID as it will be used in the Azure links. Edit your Synapse config file and change the oidc_config section:

oidc_providers:
  - idp_id: microsoft
    idp_name: Microsoft
    issuer: "https://login.microsoftonline.com/<tenant id>/v2.0"
    client_id: "<client id>"
    client_secret: "<client secret>"
    scopes: ["openid", "profile"]
    authorization_endpoint: "https://login.microsoftonline.com/<tenant id>/oauth2/v2.0/authorize"
    token_endpoint: "https://login.microsoftonline.com/<tenant id>/oauth2/v2.0/token"
    userinfo_endpoint: "https://graph.microsoft.com/oidc/userinfo"

    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username.split('@')[0] }}"
        display_name_template: "{{ user.name }}"

Dex

Dex is a simple, open-source OpenID Connect Provider. Although it is designed to help building a full-blown provider with an external database, it can be configured with static passwords in a config file.

Follow the Getting Started guide to install Dex.

Edit examples/config-dev.yaml config file from the Dex repo to add a client:

staticClients:
- id: synapse
  secret: secret
  redirectURIs:
  - '[synapse public baseurl]/_synapse/client/oidc/callback'
  name: 'Synapse'

Run with dex serve examples/config-dev.yaml.

Synapse config:

oidc_providers:
  - idp_id: dex
    idp_name: "My Dex server"
    skip_verification: true # This is needed as Dex is served on an insecure endpoint
    issuer: "http://127.0.0.1:5556/dex"
    client_id: "synapse"
    client_secret: "secret"
    scopes: ["openid", "profile"]
    user_mapping_provider:
      config:
        localpart_template: "{{ user.name }}"
        display_name_template: "{{ user.name|capitalize }}"

Keycloak

Keycloak is an opensource IdP maintained by Red Hat.

Follow the Getting Started Guide to install Keycloak and set up a realm.

1. Click Clients in the sidebar and click Create

2. Fill in the fields as below:

3. Click Save
4. Fill in the fields as below:

5. Click Save
6. On the Credentials tab, update the fields:

7. Click Regenerate Secret
8. Copy Secret

oidc_providers:
  - idp_id: keycloak
    idp_name: "My KeyCloak server"
    issuer: "https://127.0.0.1:8443/realms/{realm_name}"
    client_id: "synapse"
    client_secret: "copy secret generated from above"
    scopes: ["openid", "profile"]
    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username }}"
        display_name_template: "{{ user.name }}"

Auth0

Auth0 is a hosted SaaS IdP solution.

1. Create a regular web application for Synapse

2. Set the Allowed Callback URLs to [synapse public baseurl]/_synapse/client/oidc/callback

3. Add a rule with any name to add the preferred_username claim. (See https://auth0.com/docs/customize/rules/create-rules for more information on how to create rules.)

   Code sample

   function addPersistenceAttribute(user, context, callback) {
     user.user_metadata = user.user_metadata || {};
     user.user_metadata.preferred_username = user.user_metadata.preferred_username || user.user_id;
     context.idToken.preferred_username = user.user_metadata.preferred_username;
   
     auth0.users.updateUserMetadata(user.user_id, user.user_metadata)
       .then(function(){
           callback(null, user, context);
       })
       .catch(function(err){
           callback(err);
       });
   }
   

Synapse config:

oidc_providers:
  - idp_id: auth0
    idp_name: Auth0
    issuer: "https://your-tier.eu.auth0.com/" # TO BE FILLED
    client_id: "your-client-id" # TO BE FILLED
    client_secret: "your-client-secret" # TO BE FILLED
    scopes: ["openid", "profile"]
    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username }}"
        display_name_template: "{{ user.name }}"

Authentik

Authentik is an open-source IdP solution.

1. Create a provider in Authentik, with type OAuth2/OpenID.
2. The parameters are:

• Client Type: Confidential
• JWT Algorithm: RS256
• Scopes: OpenID, Email and Profile
• RSA Key: Select any available key
• Redirect URIs: [synapse public baseurl]/_synapse/client/oidc/callback

3. Create an application for synapse in Authentik and link it to the provider.
4. Note the slug of your application, Client ID and Client Secret.

Note: RSA keys must be used for signing for Authentik, ECC keys do not work.

Synapse config:

oidc_providers:
  - idp_id: authentik
    idp_name: authentik
    discover: true
    issuer: "https://your.authentik.example.org/application/o/your-app-slug/" # TO BE FILLED: domain and slug
    client_id: "your client id" # TO BE FILLED
    client_secret: "your client secret" # TO BE FILLED
    scopes:
      - "openid"
      - "profile"
      - "email"
    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username }}"
        display_name_template: "{{ user.preferred_username|capitalize }}" # TO BE FILLED: If your users have names in Authentik and you want those in Synapse, this should be replaced with user.name|capitalize.

LemonLDAP

LemonLDAP::NG is an open-source IdP solution.

1. Create an OpenID Connect Relying Parties in LemonLDAP::NG
2. The parameters are:

• Client ID under the basic menu of the new Relying Parties (Options > Basic > Client ID)
• Client secret (Options > Basic > Client secret)
• JWT Algorithm: RS256 within the security menu of the new Relying Parties (Options > Security > ID Token signature algorithm and Options > Security > Access Token signature algorithm)
• Scopes: OpenID, Email and Profile
• Allowed redirection addresses for login (Options > Basic > Allowed redirection addresses for login ) : [synapse public baseurl]/_synapse/client/oidc/callback

Synapse config:

oidc_providers:
  - idp_id: lemonldap
    idp_name: lemonldap
    discover: true
    issuer: "https://auth.example.org/" # TO BE FILLED: replace with your domain
    client_id: "your client id" # TO BE FILLED
    client_secret: "your client secret" # TO BE FILLED
    scopes:
      - "openid"
      - "profile"
      - "email"
    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username }}}"
        # TO BE FILLED: If your users have names in LemonLDAP::NG and you want those in Synapse, this should be replaced with user.name|capitalize or any valid filter.
        display_name_template: "{{ user.preferred_username|capitalize }}"

GitHub

GitHub is a bit special as it is not an OpenID Connect compliant provider, but just a regular OAuth2 provider.

The /user API endpoint can be used to retrieve information on the authenticated user. As the Synapse login mechanism needs an attribute to uniquely identify users, and that endpoint does not return a sub property, an alternative subject_claim has to be set.

1. Create a new OAuth application: https://github.com/settings/applications/new.
2. Set the callback URL to [synapse public baseurl]/_synapse/client/oidc/callback.

Synapse config:

oidc_providers:
  - idp_id: github
    idp_name: Github
    idp_brand: "github"  # optional: styling hint for clients
    discover: false
    issuer: "https://github.com/"
    client_id: "your-client-id" # TO BE FILLED
    client_secret: "your-client-secret" # TO BE FILLED
    authorization_endpoint: "https://github.com/login/oauth/authorize"
    token_endpoint: "https://github.com/login/oauth/access_token"
    userinfo_endpoint: "https://api.github.com/user"
    scopes: ["read:user"]
    user_mapping_provider:
      config:
        subject_claim: "id"
        localpart_template: "{{ user.login }}"
        display_name_template: "{{ user.name }}"

Google

Google is an OpenID certified authentication and authorisation provider.

1. Set up a project in the Google API Console (see documentation).
2. Add an "OAuth Client ID" for a Web Application under "Credentials".
3. Copy the Client ID and Client Secret, and add the following to your synapse config:

   oidc_providers:
     - idp_id: google
       idp_name: Google
       idp_brand: "google"  # optional: styling hint for clients
       issuer: "https://accounts.google.com/"
       client_id: "your-client-id" # TO BE FILLED
       client_secret: "your-client-secret" # TO BE FILLED
       scopes: ["openid", "profile"]
       user_mapping_provider:
         config:
           localpart_template: "{{ user.given_name|lower }}"
           display_name_template: "{{ user.name }}"
   

4. Back in the Google console, add this Authorized redirect URI: [synapse public baseurl]/_synapse/client/oidc/callback.

Twitch

1. Setup a developer account on Twitch
2. Obtain the OAuth 2.0 credentials by creating an app
3. Add this OAuth Redirect URL: [synapse public baseurl]/_synapse/client/oidc/callback

Synapse config:

oidc_providers:
  - idp_id: twitch
    idp_name: Twitch
    issuer: "https://id.twitch.tv/oauth2/"
    client_id: "your-client-id" # TO BE FILLED
    client_secret: "your-client-secret" # TO BE FILLED
    client_auth_method: "client_secret_post"
    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username }}"
        display_name_template: "{{ user.name }}"

GitLab

1. Create a new application.
2. Add the read_user and openid scopes.
3. Add this Callback URL: [synapse public baseurl]/_synapse/client/oidc/callback

Synapse config:

oidc_providers:
  - idp_id: gitlab
    idp_name: Gitlab
    idp_brand: "gitlab"  # optional: styling hint for clients
    issuer: "https://gitlab.com/"
    client_id: "your-client-id" # TO BE FILLED
    client_secret: "your-client-secret" # TO BE FILLED
    client_auth_method: "client_secret_post"
    scopes: ["openid", "read_user"]
    user_profile_method: "userinfo_endpoint"
    user_mapping_provider:
      config:
        localpart_template: '{{ user.nickname }}'
        display_name_template: '{{ user.name }}'

Facebook

0. You will need a Facebook developer account. You can register for one here.
1. On the apps page of the developer console, "Create App", and choose "Build Connected Experiences".
2. Once the app is created, add "Facebook Login" and choose "Web". You don't need to go through the whole form here.
3. In the left-hand menu, open "Products"/"Facebook Login"/"Settings".
   • Add [synapse public baseurl]/_synapse/client/oidc/callback as an OAuth Redirect URL.
4. In the left-hand menu, open "Settings/Basic". Here you can copy the "App ID" and "App Secret" for use below.

Synapse config:

  - idp_id: facebook
    idp_name: Facebook
    idp_brand: "facebook"  # optional: styling hint for clients
    discover: false
    issuer: "https://www.facebook.com"
    client_id: "your-client-id" # TO BE FILLED
    client_secret: "your-client-secret" # TO BE FILLED
    scopes: ["openid", "email"]
    authorization_endpoint: "https://facebook.com/dialog/oauth"
    token_endpoint: "https://graph.facebook.com/v9.0/oauth/access_token"
    jwks_uri: "https://www.facebook.com/.well-known/oauth/openid/jwks/"
    user_mapping_provider:
      config:
        display_name_template: "{{ user.name }}"
        email_template: "{{ '{{ user.email }}' }}"

Relevant documents:

• Manually Build a Login Flow
• Using Facebook's Graph API
• Reference to the User endpoint

Facebook do have an OIDC discovery endpoint, but it has a response_types_supported which excludes "code" (which we rely on, and is even mentioned in their documentation), so we have to disable discovery and configure the URIs manually.

Gitea

Gitea is, like Github, not an OpenID provider, but just an OAuth2 provider.

The /user API endpoint can be used to retrieve information on the authenticated user. As the Synapse login mechanism needs an attribute to uniquely identify users, and that endpoint does not return a sub property, an alternative subject_claim has to be set.

1. Create a new application.
2. Add this Callback URL: [synapse public baseurl]/_synapse/client/oidc/callback

Synapse config:

oidc_providers:
  - idp_id: gitea
    idp_name: Gitea
    discover: false
    issuer: "https://your-gitea.com/"
    client_id: "your-client-id" # TO BE FILLED
    client_secret: "your-client-secret" # TO BE FILLED
    client_auth_method: client_secret_post
    scopes: [] # Gitea doesn't support Scopes
    authorization_endpoint: "https://your-gitea.com/login/oauth/authorize"
    token_endpoint: "https://your-gitea.com/login/oauth/access_token"
    userinfo_endpoint: "https://your-gitea.com/api/v1/user"
    user_mapping_provider:
      config:
        subject_claim: "id"
        localpart_template: "{{ user.login }}"
        display_name_template: "{{ user.full_name }}"

XWiki

Install OpenID Connect Provider extension in your XWiki instance.

Synapse config:

oidc_providers:
  - idp_id: xwiki
    idp_name: "XWiki"
    issuer: "https://myxwikihost/xwiki/oidc/"
    client_id: "your-client-id" # TO BE FILLED
    client_auth_method: none
    scopes: ["openid", "profile"]
    user_profile_method: "userinfo_endpoint"
    user_mapping_provider:
      config:
        localpart_template: "{{ user.preferred_username }}"
        display_name_template: "{{ user.name }}"

Apple

Configuring "Sign in with Apple" (SiWA) requires an Apple Developer account.

You will need to create a new "Services ID" for SiWA, and create and download a private key with "SiWA" enabled.

As well as the private key file, you will need:

• Client ID: the "identifier" you gave the "Services ID"
• Team ID: a 10-character ID associated with your developer account.
• Key ID: the 10-character identifier for the key.

Apple's developer documentation has more information on setting up SiWA.

The synapse config will look like this:

  - idp_id: apple
    idp_name: Apple
    issuer: "https://appleid.apple.com"
    client_id: "your-client-id" # Set to the "identifier" for your "ServicesID"
    client_auth_method: "client_secret_post"
    client_secret_jwt_key:
      key_file: "/path/to/AuthKey_KEYIDCODE.p8"  # point to your key file
      jwt_header:
        alg: ES256
        kid: "KEYIDCODE"   # Set to the 10-char Key ID
      jwt_payload:
        iss: TEAMIDCODE    # Set to the 10-char Team ID
    scopes: ["name", "email", "openid"]
    authorization_endpoint: https://appleid.apple.com/auth/authorize?response_mode=form_post
    user_mapping_provider:
      config:
        email_template: "{{ user.email }}"

Django OAuth Toolkit

django-oauth-toolkit is a Django application providing out of the box all the endpoints, data and logic needed to add OAuth2 capabilities to your Django projects. It supports OpenID Connect too.

Configuration on Django's side:

1. Add an application: https://example.com/admin/oauth2_provider/application/add/ and choose parameters like this:

• Redirect uris: https://synapse.example.com/_synapse/client/oidc/callback
• Client type: Confidential
• Authorization grant type: Authorization code
• Algorithm: HMAC with SHA-2 256

2. You can customize the claims Django gives to synapse (optional):

   Code sample

   class CustomOAuth2Validator(OAuth2Validator):
   
       def get_additional_claims(self, request):
           return {
               "sub": request.user.email,
               "email": request.user.email,
               "first_name": request.user.first_name,
               "last_name": request.user.last_name,
           }
   

Your synapse config is then:

oidc_providers:
  - idp_id: django_example
    idp_name: "Django Example"
    issuer: "https://example.com/o/"
    client_id: "your-client-id"  # CHANGE ME
    client_secret: "your-client-secret"  # CHANGE ME
    scopes: ["openid"]
    user_profile_method: "userinfo_endpoint"  # needed because oauth-toolkit does not include user information in the authorization response
    user_mapping_provider:
      config:
        localpart_template: "{{ user.email.split('@')[0] }}"
        display_name_template: "{{ user.first_name }} {{ user.last_name }}"
        email_template: "{{ user.email }}"

SAML

Synapse supports authenticating users via the Security Assertion Markup Language (SAML) protocol natively.

Please see the saml2_config and sso sections of the Synapse configuration file for more details.

CAS

Synapse supports authenticating users via the Central Authentication Service protocol (CAS) natively.

Please see the cas_config and sso sections of the configuration manual for more details.

SSO Mapping Providers

A mapping provider is a Python class (loaded via a Python module) that works out how to map attributes of a SSO response to Matrix-specific user attributes. Details such as user ID localpart, displayname, and even avatar URLs are all things that can be mapped from talking to a SSO service.

As an example, a SSO service may return the email address "john.smith@example.com" for a user, whereas Synapse will need to figure out how to turn that into a displayname when creating a Matrix user for this individual. It may choose John Smith, or Smith, John [Example.com] or any number of variations. As each Synapse configuration may want something different, this is where SAML mapping providers come into play.

SSO mapping providers are currently supported for OpenID and SAML SSO configurations. Please see the details below for how to implement your own.

It is up to the mapping provider whether the user should be assigned a predefined Matrix ID based on the SSO attributes, or if the user should be allowed to choose their own username.

In the first case - where users are automatically allocated a Matrix ID - it is the responsibility of the mapping provider to normalise the SSO attributes and map them to a valid Matrix ID. The specification for Matrix IDs has some information about what is considered valid.

If the mapping provider does not assign a Matrix ID, then Synapse will automatically serve an HTML page allowing the user to pick their own username.

External mapping providers are provided to Synapse in the form of an external Python module. You can retrieve this module from PyPI or elsewhere, but it must be importable via Synapse (e.g. it must be in the same virtualenv as Synapse). The Synapse config is then modified to point to the mapping provider (and optionally provide additional configuration for it).

OpenID Mapping Providers

The OpenID mapping provider can be customized by editing the oidc_providers.user_mapping_provider.module config option.

oidc_providers.user_mapping_provider.config allows you to provide custom configuration options to the module. Check with the module's documentation for what options it provides (if any). The options listed by default are for the user mapping provider built in to Synapse. If using a custom module, you should comment these options out and use those specified by the module instead.

Building a Custom OpenID Mapping Provider

A custom mapping provider must specify the following methods:

• def __init__(self, parsed_config)
  • Arguments:
    • parsed_config - A configuration object that is the return value of the parse_config method. You should set any configuration options needed by the module here.
• def parse_config(config)
  • This method should have the @staticmethod decoration.
  • Arguments:
    • config - A dict representing the parsed content of the oidc_providers.user_mapping_provider.config homeserver config option. Runs on homeserver startup. Providers should extract and validate any option values they need here.
  • Whatever is returned will be passed back to the user mapping provider module's __init__ method during construction.
• def get_remote_user_id(self, userinfo)
  • Arguments:
    • userinfo - A authlib.oidc.core.claims.UserInfo object to extract user information from.
  • This method must return a string, which is the unique, immutable identifier for the user. Commonly the sub claim of the response.
• async def map_user_attributes(self, userinfo, token, failures)
  • This method must be async.
  • Arguments:
    • userinfo - A authlib.oidc.core.claims.UserInfo object to extract user information from.
    • token - A dictionary which includes information necessary to make further requests to the OpenID provider.
    • failures - An int that represents the amount of times the returned mxid localpart mapping has failed. This should be used to create a deduplicated mxid localpart which should be returned instead. For example, if this method returns john.doe as the value of localpart in the returned dict, and that is already taken on the homeserver, this method will be called again with the same parameters but with failures=1. The method should then return a different localpart value, such as john.doe1.
  • Returns a dictionary with two keys:
    • localpart: A string, used to generate the Matrix ID. If this is None, the user is prompted to pick their own username. This is only used during a user's first login. Once a localpart has been associated with a remote user ID (see get_remote_user_id) it cannot be updated.
    • displayname: An optional string, the display name for the user.
• async def get_extra_attributes(self, userinfo, token)

  • This method must be async.

  • Arguments:

    • userinfo - A authlib.oidc.core.claims.UserInfo object to extract user information from.
    • token - A dictionary which includes information necessary to make further requests to the OpenID provider.

  • Returns a dictionary that is suitable to be serialized to JSON. This will be returned as part of the response during a successful login.

    Note that care should be taken to not overwrite any of the parameters usually returned as part of the login response.

Default OpenID Mapping Provider

Synapse has a built-in OpenID mapping provider if a custom provider isn't specified in the config. It is located at synapse.handlers.oidc.JinjaOidcMappingProvider.

SAML Mapping Providers

The SAML mapping provider can be customized by editing the saml2_config.user_mapping_provider.module config option.

saml2_config.user_mapping_provider.config allows you to provide custom configuration options to the module. Check with the module's documentation for what options it provides (if any). The options listed by default are for the user mapping provider built in to Synapse. If using a custom module, you should comment these options out and use those specified by the module instead.

Building a Custom SAML Mapping Provider

A custom mapping provider must specify the following methods:

• def __init__(self, parsed_config, module_api)
  • Arguments:
    • parsed_config - A configuration object that is the return value of the parse_config method. You should set any configuration options needed by the module here.
    • module_api - a synapse.module_api.ModuleApi object which provides the stable API available for extension modules.
• def parse_config(config)
  • This method should have the @staticmethod decoration.
  • Arguments:
    • config - A dict representing the parsed content of the saml_config.user_mapping_provider.config homeserver config option. Runs on homeserver startup. Providers should extract and validate any option values they need here.
  • Whatever is returned will be passed back to the user mapping provider module's __init__ method during construction.
• def get_saml_attributes(config)
  • This method should have the @staticmethod decoration.
  • Arguments:
    • config - A object resulting from a call to parse_config.
  • Returns a tuple of two sets. The first set equates to the SAML auth response attributes that are required for the module to function, whereas the second set consists of those attributes which can be used if available, but are not necessary.
• def get_remote_user_id(self, saml_response, client_redirect_url)
  • Arguments:
    • saml_response - A saml2.response.AuthnResponse object to extract user information from.
    • client_redirect_url - A string, the URL that the client will be redirected to.
  • This method must return a string, which is the unique, immutable identifier for the user. Commonly the uid claim of the response.
• def saml_response_to_user_attributes(self, saml_response, failures, client_redirect_url)

  • Arguments:

    • saml_response - A saml2.response.AuthnResponse object to extract user information from.
    • failures - An int that represents the amount of times the returned mxid localpart mapping has failed. This should be used to create a deduplicated mxid localpart which should be returned instead. For example, if this method returns john.doe as the value of mxid_localpart in the returned dict, and that is already taken on the homeserver, this method will be called again with the same parameters but with failures=1. The method should then return a different mxid_localpart value, such as john.doe1.
    • client_redirect_url - A string, the URL that the client will be redirected to.

  • This method must return a dictionary, which will then be used by Synapse to build a new user. The following keys are allowed:

    • mxid_localpart - A string, the mxid localpart of the new user. If this is None, the user is prompted to pick their own username. This is only used during a user's first login. Once a localpart has been associated with a remote user ID (see get_remote_user_id) it cannot be updated.
    • displayname - The displayname of the new user. If not provided, will default to the value of mxid_localpart.
    • emails - A list of emails for the new user. If not provided, will default to an empty list.

    Alternatively it can raise a synapse.api.errors.RedirectException to redirect the user to another page. This is useful to prompt the user for additional information, e.g. if you want them to provide their own username. It is the responsibility of the mapping provider to either redirect back to client_redirect_url (including any additional information) or to complete registration using methods from the ModuleApi.

Default SAML Mapping Provider

Synapse has a built-in SAML mapping provider if a custom provider isn't specified in the config. It is located at synapse.handlers.saml.DefaultSamlMappingProvider.

This page of the Synapse documentation is now deprecated. For up to date documentation on setting up or writing a password auth provider module, please see this page.

Password auth provider modules

Password auth providers offer a way for server administrators to integrate their Synapse installation with an existing authentication system.

A password auth provider is a Python class which is dynamically loaded into Synapse, and provides a number of methods by which it can integrate with the authentication system.

This document serves as a reference for those looking to implement their own password auth providers. Additionally, here is a list of known password auth provider module implementations:

• matrix-synapse-ldap3
• matrix-synapse-shared-secret-auth
• matrix-synapse-rest-password-provider

Required methods

Password auth provider classes must provide the following methods:

• parse_config(config) This method is passed the config object for this module from the homeserver configuration file.

  It should perform any appropriate sanity checks on the provided configuration, and return an object which is then passed into __init__.

  This method should have the @staticmethod decoration.

• __init__(self, config, account_handler)

  The constructor is passed the config object returned by parse_config, and a synapse.module_api.ModuleApi object which allows the password provider to check if accounts exist and/or create new ones.

Optional methods

Password auth provider classes may optionally provide the following methods:

• get_db_schema_files(self)

  This method, if implemented, should return an Iterable of (name, stream) pairs of database schema files. Each file is applied in turn at initialisation, and a record is then made in the database so that it is not re-applied on the next start.

• get_supported_login_types(self)

  This method, if implemented, should return a dict mapping from a login type identifier (such as m.login.password) to an iterable giving the fields which must be provided by the user in the submission to the /login API. These fields are passed in the login_dict dictionary to check_auth.

  For example, if a password auth provider wants to implement a custom login type of com.example.custom_login, where the client is expected to pass the fields secret1 and secret2, the provider should implement this method and return the following dict:

  {"com.example.custom_login": ("secret1", "secret2")}
  

• check_auth(self, username, login_type, login_dict)

  This method does the real work. If implemented, it will be called for each login attempt where the login type matches one of the keys returned by get_supported_login_types.

  It is passed the (possibly unqualified) user field provided by the client, the login type, and a dictionary of login secrets passed by the client.

  The method should return an Awaitable object, which resolves to the canonical @localpart:domain user ID if authentication is successful, and None if not.

  Alternatively, the Awaitable can resolve to a (str, func) tuple, in which case the second field is a callback which will be called with the result from the /login call (including access_token, device_id, etc.)

• check_3pid_auth(self, medium, address, password)

  This method, if implemented, is called when a user attempts to register or log in with a third party identifier, such as email. It is passed the medium (ex. "email"), an address (ex. "jdoe@example.com") and the user's password.

  The method should return an Awaitable object, which resolves to a str containing the user's (canonical) User id if authentication was successful, and None if not.

  As with check_auth, the Awaitable may alternatively resolve to a (user_id, callback) tuple.

• check_password(self, user_id, password)

  This method provides a simpler interface than get_supported_login_types and check_auth for password auth providers that just want to provide a mechanism for validating m.login.password logins.

  If implemented, it will be called to check logins with an m.login.password login type. It is passed a qualified @localpart:domain user id, and the password provided by the user.

  The method should return an Awaitable object, which resolves to True if authentication is successful, and False if not.

• on_logged_out(self, user_id, device_id, access_token)

  This method, if implemented, is called when a user logs out. It is passed the qualified user ID, the ID of the deactivated device (if any: access tokens are occasionally created without an associated device ID), and the (now deactivated) access token.

  It may return an Awaitable object; the logout request will wait for the Awaitable to complete, but the result is ignored.

JWT Login Type

Synapse comes with a non-standard login type to support JSON Web Tokens. In general the documentation for the login endpoint is still valid (and the mechanism works similarly to the token based login).

To log in using a JSON Web Token, clients should submit a /login request as follows:

{
  "type": "org.matrix.login.jwt",
  "token": "<jwt>"
}

The token field should include the JSON web token with the following claims:

• A claim that encodes the local part of the user ID is required. By default, the sub (subject) claim is used, or a custom claim can be set in the configuration file.
• The expiration time (exp), not before time (nbf), and issued at (iat) claims are optional, but validated if present.
• The issuer (iss) claim is optional, but required and validated if configured.
• The audience (aud) claim is optional, but required and validated if configured. Providing the audience claim when not configured will cause validation to fail.

In the case that the token is not valid, the homeserver must respond with 403 Forbidden and an error code of M_FORBIDDEN.

As with other login types, there are additional fields (e.g. device_id and initial_device_display_name) which can be included in the above request.

Preparing Synapse

The JSON Web Token integration in Synapse uses the Authlib library, which must be installed as follows:

• The relevant libraries are included in the Docker images and Debian packages provided by matrix.org so no further action is needed.

• If you installed Synapse into a virtualenv, run /path/to/env/bin/pip install synapse[jwt] to install the necessary dependencies.

• For other installation mechanisms, see the documentation provided by the maintainer.

To enable the JSON web token integration, you should then add a jwt_config option to your configuration file. See the configuration manual for some sample settings.

How to test JWT as a developer

Although JSON Web Tokens are typically generated from an external server, the example below uses a locally generated JWT.

1. Configure Synapse with JWT logins, note that this example uses a pre-shared secret and an algorithm of HS256:

   jwt_config:
       enabled: true
       secret: "my-secret-token"
       algorithm: "HS256"
   

2. Generate a JSON web token:

   You can use the following short Python snippet to generate a JWT protected by an HMAC. Take care that the secret and the algorithm given in the header match the entries from jwt_config above.

   from authlib.jose import jwt
   
   header = {"alg": "HS256"}
   payload = {"sub": "user1", "aud": ["audience"]}
   secret = "my-secret-token"
   result = jwt.encode(header, payload, secret)
   print(result.decode("ascii"))
   

3. Query for the login types and ensure org.matrix.login.jwt is there:

   curl http://localhost:8080/_matrix/client/r0/login
   

4. Login used the generated JSON web token from above:

   $ curl http://localhost:8082/_matrix/client/r0/login -X POST \
       --data '{"type":"org.matrix.login.jwt","token":"eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJzdWIiOiJ0ZXN0LXVzZXIifQ.Ag71GT8v01UO3w80aqRPTeuVPBIBZkYhNTJJ-_-zQIc"}'
   {
       "access_token": "<access token>",
       "device_id": "ACBDEFGHI",
       "home_server": "localhost:8080",
       "user_id": "@test-user:localhost:8480"
   }
   

You should now be able to use the returned access token to query the client API.

Refresh Tokens

Synapse supports refresh tokens since version 1.49 (some earlier versions had support for an earlier, experimental draft of MSC2918 which is not compatible).

Background and motivation

Synapse users' sessions are identified by access tokens; access tokens are issued to users on login. Each session gets a unique access token which identifies it; the access token must be kept secret as it grants access to the user's account.

Traditionally, these access tokens were eternally valid (at least until the user explicitly chose to log out).

In some cases, it may be desirable for these access tokens to expire so that the potential damage caused by leaking an access token is reduced. On the other hand, forcing a user to re-authenticate (log in again) often might be too much of an inconvenience.

Refresh tokens are a mechanism to avoid some of this inconvenience whilst still getting most of the benefits of short access token lifetimes. Refresh tokens are also a concept present in OAuth 2 — further reading is available here.

When refresh tokens are in use, both an access token and a refresh token will be issued to users on login. The access token will expire after a predetermined amount of time, but otherwise works in the same way as before. When the access token is close to expiring (or has expired), the user's client should present the homeserver (Synapse) with the refresh token.

The homeserver will then generate a new access token and refresh token for the user and return them. The old refresh token is invalidated and can not be used again*.

Finally, refresh tokens also make it possible for sessions to be logged out if they are inactive for too long, before the session naturally ends; see the configuration guide below.

*To prevent issues if clients lose connection half-way through refreshing a token, the refresh token is only invalidated once the new access token has been used at least once. For all intents and purposes, the above simplification is sufficient.

Caveats

There are some caveats:

• If a third party gets both your access token and refresh token, they will be able to continue to enjoy access to your session.
  • This is still an improvement because you (the user) will notice when your session expires and you're not able to use your refresh token. That would be a giveaway that someone else has compromised your session. You would be able to log in again and terminate that session. Previously (with long-lived access tokens), a third party that has your access token could go undetected for a very long time.
• Clients need to implement support for refresh tokens in order for them to be a useful mechanism.
  • It is up to homeserver administrators if they want to issue long-lived access tokens to clients not implementing refresh tokens.
    • For compatibility, it is likely that they should, at least until client support is widespread.
      • Users with clients that support refresh tokens will still benefit from the added security; it's not possible to downgrade a session to using long-lived access tokens so this effectively gives users the choice.
    • In a closed environment where all users use known clients, this may not be an issue as the homeserver administrator can know if the clients have refresh token support. In that case, the non-refreshable access token lifetime may be set to a short duration so that a similar level of security is provided.

Configuration Guide

The following configuration options, in the registration section, are related:

• session_lifetime: maximum length of a session, even if it's refreshed. In other words, the client must log in again after this time period. In most cases, this can be unset (infinite) or set to a long time (years or months).
• refreshable_access_token_lifetime: lifetime of access tokens that are created by clients supporting refresh tokens. This should be short; a good value might be 5 minutes (5m).
• nonrefreshable_access_token_lifetime: lifetime of access tokens that are created by clients which don't support refresh tokens. Make this short if you want to effectively force use of refresh tokens. Make this long if you don't want to inconvenience users of clients which don't support refresh tokens (by forcing them to frequently re-authenticate using login credentials).
• refresh_token_lifetime: lifetime of refresh tokens. In other words, the client must refresh within this time period to maintain its session. Unless you want to log inactive sessions out, it is often fine to use a long value here or even leave it unset (infinite). Beware that making it too short will inconvenience clients that do not connect very often, including mobile clients and clients of infrequent users (by making it more difficult for them to refresh in time, which may force them to need to re-authenticate using login credentials).

Note: All four options above only apply when tokens are created (by logging in or refreshing). Changes to these settings do not apply retroactively.

Using refresh token expiry to log out inactive sessions

If you'd like to force sessions to be logged out upon inactivity, you can enable refreshable access token expiry and refresh token expiry.

This works because a client must refresh at least once within a period of refresh_token_lifetime in order to maintain valid credentials to access the account.

(It's suggested that refresh_token_lifetime should be longer than refreshable_access_token_lifetime and this section assumes that to be the case for simplicity.)

Note: this will only affect sessions using refresh tokens. You may wish to set a short nonrefreshable_access_token_lifetime to prevent this being bypassed by clients that do not support refresh tokens.

Choosing values that guarantee permitting some inactivity

It may be desirable to permit some short periods of inactivity, for example to accommodate brief outages in client connectivity.

The following model aims to provide guidance for choosing refresh_token_lifetime and refreshable_access_token_lifetime to satisfy requirements of the form:

1. inactivity longer than L MUST cause the session to be logged out; and
2. inactivity shorter than S MUST NOT cause the session to be logged out.

This model makes the weakest assumption that all active clients will refresh as needed to maintain an active access token, but no sooner. In reality, clients may refresh more often than this model assumes, but the above requirements will still hold.

To satisfy the above model,

• refresh_token_lifetime should be set to L; and
• refreshable_access_token_lifetime should be set to L - S.

Overview

A captcha can be enabled on your homeserver to help prevent bots from registering accounts. Synapse currently uses Google's reCAPTCHA service which requires API keys from Google.

Getting API keys

1. Create a new site at https://www.google.com/recaptcha/admin/create
2. Set the label to anything you want
3. Set the type to reCAPTCHA v2 using the "I'm not a robot" Checkbox option. This is the only type of captcha that works with Synapse.
4. Add the public hostname for your server, as set in public_baseurl in homeserver.yaml, to the list of authorized domains. If you have not set public_baseurl, use server_name.
5. Agree to the terms of service and submit.
6. Copy your site key and secret key and add them to your homeserver.yaml configuration file

   recaptcha_public_key: YOUR_SITE_KEY
   recaptcha_private_key: YOUR_SECRET_KEY
   

7. Enable the CAPTCHA for new registrations

   enable_registration_captcha: true
   

8. Go to the settings page for the CAPTCHA you just created
9. Uncheck the "Verify the origin of reCAPTCHA solutions" checkbox so that the captcha can be displayed in any client. If you do not disable this option then you must specify the domains of every client that is allowed to display the CAPTCHA.

Configuring IP used for auth

The reCAPTCHA API requires that the IP address of the user who solved the CAPTCHA is sent. If the client is connecting through a proxy or load balancer, it may be required to use the X-Forwarded-For (XFF) header instead of the origin IP address. This can be configured using the x_forwarded directive in the listeners section of the homeserver.yaml configuration file.

Registering an Application Service

The registration of new application services depends on the homeserver used. In synapse, you need to create a new configuration file for your AS and add it to the list specified under the app_service_config_files config option in your synapse config.

For example:

app_service_config_files:
- /home/matrix/.synapse/<your-AS>.yaml

The format of the AS configuration file is as follows:

url: <base url of AS>
as_token: <token AS will add to requests to HS>
hs_token: <token HS will add to requests to AS>
sender_localpart: <localpart of AS user>
namespaces:
  users:  # List of users we're interested in
    - exclusive: <bool>
      regex: <regex>
      group_id: <group>
    - ...
  aliases: []  # List of aliases we're interested in
  rooms: [] # List of room ids we're interested in

exclusive: If enabled, only this application service is allowed to register users in its namespace(s). group_id: All users of this application service are dynamically joined to this group. This is useful for e.g user organisation or flairs.

See the spec for further details on how application services work.

Server Notices

'Server Notices' are a new feature introduced in Synapse 0.30. They provide a channel whereby server administrators can send messages to users on the server.

They are used as part of communication of the server polices (see Consent Tracking), however the intention is that they may also find a use for features such as "Message of the day".

This is a feature specific to Synapse, but it uses standard Matrix communication mechanisms, so should work with any Matrix client.

User experience

When the user is first sent a server notice, they will get an invitation to a room (typically called 'Server Notices', though this is configurable in homeserver.yaml). They will be unable to reject this invitation - attempts to do so will receive an error.

Once they accept the invitation, they will see the notice message in the room history; it will appear to have come from the 'server notices user' (see below).

The user is prevented from sending any messages in this room by the power levels.

Having joined the room, the user can leave the room if they want. Subsequent server notices will then cause a new room to be created.

Synapse configuration

Server notices come from a specific user id on the server. Server administrators are free to choose the user id - something like server is suggested, meaning the notices will come from @server:<your_server_name>. Once the Server Notices user is configured, that user id becomes a special, privileged user, so administrators should ensure that it is not already allocated.

In order to support server notices, it is necessary to add some configuration to the homeserver.yaml file. In particular, you should add a server_notices section, which should look like this:

server_notices:
   system_mxid_localpart: server
   system_mxid_display_name: "Server Notices"
   system_mxid_avatar_url: "mxc://server.com/oumMVlgDnLYFaPVkExemNVVZ"
   room_name: "Server Notices"

The only compulsory setting is system_mxid_localpart, which defines the user id of the Server Notices user, as above. room_name defines the name of the room which will be created.

system_mxid_display_name and system_mxid_avatar_url can be used to set the displayname and avatar of the Server Notices user.

Sending notices

To send server notices to users you can use the admin_api.

Support in Synapse for tracking agreement to server terms and conditions

Synapse 0.30 introduces support for tracking whether users have agreed to the terms and conditions set by the administrator of a server - and blocking access to the server until they have.

There are several parts to this functionality; each requires some specific configuration in homeserver.yaml to be enabled.

Note that various parts of the configuation and this document refer to the "privacy policy": agreement with a privacy policy is one particular use of this feature, but of course adminstrators can specify other terms and conditions unrelated to "privacy" per se.

Collecting policy agreement from a user

Synapse can be configured to serve the user a simple policy form with an "accept" button. Clicking "Accept" records the user's acceptance in the database and shows a success page.

To enable this, first create templates for the policy and success pages. These should be stored on the local filesystem.

These templates use the Jinja2 templating language, and docs/privacy_policy_templates gives examples of the sort of thing that can be done.

Note that the templates must be stored under a name giving the language of the template - currently this must always be en (for "English"); internationalisation support is intended for the future.

The template for the policy itself should be versioned and named according to the version: for example 1.0.html. The version of the policy which the user has agreed to is stored in the database.

Once the templates are in place, make the following changes to homeserver.yaml:

1. Add a user_consent section, which should look like:

   user_consent:
     template_dir: privacy_policy_templates
     version: 1.0
   

   template_dir points to the directory containing the policy templates. version defines the version of the policy which will be served to the user. In the example above, Synapse will serve privacy_policy_templates/en/1.0.html.

2. Add a form_secret setting at the top level:

   form_secret: "<unique secret>"
   

   This should be set to an arbitrary secret string (try pwgen -y 30 to generate suitable secrets).

   More on what this is used for below.

3. Add consent wherever the client resource is currently enabled in the listeners configuration. For example:

   listeners:
     - port: 8008
       resources:
         - names:
           - client
           - consent
   

Finally, ensure that jinja2 is installed. If you are using a virtualenv, this should be a matter of pip install Jinja2. On debian, try apt-get install python-jinja2.

Once this is complete, and the server has been restarted, try visiting https://<server>/_matrix/consent. If correctly configured, this should give an error "Missing string query parameter 'u'". It is now possible to manually construct URIs where users can give their consent.

Enabling consent tracking at registration

1. Add the following to your configuration:

   user_consent:
     require_at_registration: true
     policy_name: "Privacy Policy" # or whatever you'd like to call the policy
   

2. In your consent templates, make use of the public_version variable to see if an unauthenticated user is viewing the page. This is typically wrapped around the form that would be used to actually agree to the document:

   {% if not public_version %}
     <!-- The variables used here are only provided when the 'u' param is given to the homeserver -->
     <form method="post" action="consent">
       <input type="hidden" name="v" value="{{version}}"/>
       <input type="hidden" name="u" value="{{user}}"/>
       <input type="hidden" name="h" value="{{userhmac}}"/>
       <input type="submit" value="Sure thing!"/>
     </form>
   {% endif %}
   

3. Restart Synapse to apply the changes.

Visiting https://<server>/_matrix/consent should now give you a view of the privacy document. This is what users will be able to see when registering for accounts.

Constructing the consent URI

It may be useful to manually construct the "consent URI" for a given user - for instance, in order to send them an email asking them to consent. To do this, take the base https://<server>/_matrix/consent URL and add the following query parameters:

• u: the user id of the user. This can either be a full MXID (@user:server.com) or just the localpart (user).

• h: hex-encoded HMAC-SHA256 of u using the form_secret as a key. It is possible to calculate this on the commandline with something like:

  echo -n '<user>' | openssl sha256 -hmac '<form_secret>'
  

  This should result in a URI which looks something like: https://<server>/_matrix/consent?u=<user>&h=68a152465a4d....

Note that not providing a u parameter will be interpreted as wanting to view the document from an unauthenticated perspective, such as prior to registration. Therefore, the h parameter is not required in this scenario. To enable this behaviour, set require_at_registration to true in your user_consent config.

Sending users a server notice asking them to agree to the policy

It is possible to configure Synapse to send a server notice to anybody who has not yet agreed to the current version of the policy. To do so:

• ensure that the consent resource is configured, as in the previous section

• ensure that server notices are configured, as in the server notice documentation.

• Add server_notice_content under user_consent in homeserver.yaml. For example:

  user_consent:
    server_notice_content:
      msgtype: m.text
      body: >-
        Please give your consent to the privacy policy at %(consent_uri)s.
  

  Synapse automatically replaces the placeholder %(consent_uri)s with the consent uri for that user.

• ensure that public_baseurl is set in homeserver.yaml, and gives the base URI that clients use to connect to the server. (It is used to construct consent_uri in the server notice.)

Blocking users from using the server until they agree to the policy

Synapse can be configured to block any attempts to join rooms or send messages until the user has given their agreement to the policy. (Joining the server notices room is exempted from this).

To enable this, add block_events_error under user_consent. For example:

user_consent:
  block_events_error: >-
    You can't send any messages until you consent to the privacy policy at
    %(consent_uri)s.

Synapse automatically replaces the placeholder %(consent_uri)s with the consent uri for that user.

ensure that public_baseurl is set in homeserver.yaml, and gives the base URI that clients use to connect to the server. (It is used to construct consent_uri in the error.)

User Directory API Implementation

The user directory is currently maintained based on the 'visible' users on this particular server - i.e. ones which your account shares a room with, or who are present in a publicly viewable room present on the server.

The directory info is stored in various tables, which can (typically after DB corruption) get stale or out of sync. If this happens, for now the solution to fix it is to use the admin API and execute the job regenerate_directory. This should then start a background task to flush the current tables and regenerate the directory.

Data model

There are five relevant tables that collectively form the "user directory". Three of them track a master list of all the users we could search for. The last two (collectively called the "search tables") track who can see who.

From all of these tables we exclude three types of local user:

• support users
• appservice users
• deactivated users

• user_directory. This contains the user_id, display name and avatar we'll return when you search the directory.

  • Because there's only one directory entry per user, it's important that we only ever put publicly visible names here. Otherwise we might leak a private nickname or avatar used in a private room.
  • Indexed on rooms. Indexed on users.

• user_directory_search. To be joined to user_directory. It contains an extra column that enables full text search based on user ids and display names. Different schemas for SQLite and Postgres with different code paths to match.

  • Indexed on the full text search data. Indexed on users.

• user_directory_stream_pos. When the initial background update to populate the directory is complete, we record a stream position here. This indicates that synapse should now listen for room changes and incrementally update the directory where necessary.

• users_in_public_rooms. Contains associations between users and the public rooms they're in. Used to determine which users are in public rooms and should be publicly visible in the directory.

• users_who_share_private_rooms. Rows are triples (L, M, room id) where L is a local user and M is a local or remote user. L and M should be different, but this isn't enforced by a constraint.

Message retention policies

Synapse admins can enable support for message retention policies on their homeserver. Message retention policies exist at a room level, follow the semantics described in MSC1763, and allow server and room admins to configure how long messages should be kept in a homeserver's database before being purged from it. Please note that, as this feature isn't part of the Matrix specification yet, this implementation is to be considered as experimental. There are known bugs which may cause database corruption. Proceed with caution.

A message retention policy is mainly defined by its max_lifetime parameter, which defines how long a message can be kept around after it was sent to the room. If a room doesn't have a message retention policy, and there's no default one for a given server, then no message sent in that room is ever purged on that server.

MSC1763 also specifies semantics for a min_lifetime parameter which defines the amount of time after which an event can get purged (after it was sent to the room), but Synapse doesn't currently support it beyond registering it.

Both max_lifetime and min_lifetime are optional parameters.

Note that message retention policies don't apply to state events.

Once an event reaches its expiry date (defined as the time it was sent plus the value for max_lifetime in the room), two things happen:

• Synapse stops serving the event to clients via any endpoint.
• The message gets picked up by the next purge job (see the "Purge jobs" section) and is removed from Synapse's database.

Since purge jobs don't run continuously, this means that an event might stay in a server's database for longer than the value for max_lifetime in the room would allow, though hidden from clients.

Similarly, if a server (with support for message retention policies enabled) receives from another server an event that should have been purged according to its room's policy, then the receiving server will process and store that event until it's picked up by the next purge job, though it will always hide it from clients.

Synapse requires at least one message in each room, so it will never delete the last message in a room. It will, however, hide it from clients.

Server configuration

Support for this feature can be enabled and configured by adding a the retention in the Synapse configuration file (see configuration manual).

To enable support for message retention policies, set the setting enabled in this section to true.

Default policy

A default message retention policy is a policy defined in Synapse's configuration that is used by Synapse for every room that doesn't have a message retention policy configured in its state. This allows server admins to ensure that messages are never kept indefinitely in a server's database.

A default policy can be defined as such, by adding the retention option in the configuration file and adding these sub-options:

default_policy:
  min_lifetime: 1d
  max_lifetime: 1y

Here, min_lifetime and max_lifetime have the same meaning and level of support as previously described. They can be expressed either as a duration (using the units s (seconds), m (minutes), h (hours), d (days), w (weeks) and y (years)) or as a number of milliseconds.

Purge jobs

Purge jobs are the jobs that Synapse runs in the background to purge expired events from the database. They are only run if support for message retention policies is enabled in the server's configuration. If no configuration for purge jobs is configured by the server admin, Synapse will use a default configuration, which is described here in the configuration manual.

Some server admins might want a finer control on when events are removed depending on an event's room's policy. This can be done by setting the purge_jobs sub-section in the retention section of the configuration file. An example of such configuration could be:

purge_jobs:
  - longest_max_lifetime: 3d
    interval: 12h
  - shortest_max_lifetime: 3d
    longest_max_lifetime: 1w
    interval: 1d
  - shortest_max_lifetime: 1w
    interval: 2d

In this example, we define three jobs:

• one that runs twice a day (every 12 hours) and purges events in rooms which policy's max_lifetime is lower or equal to 3 days.
• one that runs once a day and purges events in rooms which policy's max_lifetime is between 3 days and a week.
• one that runs once every 2 days and purges events in rooms which policy's max_lifetime is greater than a week.

Note that this example is tailored to show different configurations and features slightly more jobs than it's probably necessary (in practice, a server admin would probably consider it better to replace the two last jobs with one that runs once a day and handles rooms which policy's max_lifetime is greater than 3 days).

Keep in mind, when configuring these jobs, that a purge job can become quite heavy on the server if it targets many rooms, therefore prefer having jobs with a low interval that target a limited set of rooms. Also make sure to include a job with no minimum and one with no maximum to make sure your configuration handles every policy.

As previously mentioned in this documentation, while a purge job that runs e.g. every day means that an expired event might stay in the database for up to a day after its expiry, Synapse hides expired events from clients as soon as they expire, so the event is not visible to local users between its expiry date and the moment it gets purged from the server's database.

Lifetime limits

Server admins can set limits on the values of max_lifetime to use when purging old events in a room. These limits can be defined under the retention option in the configuration file:

allowed_lifetime_min: 1d
allowed_lifetime_max: 1y

The limits are considered when running purge jobs. If necessary, the effective value of max_lifetime will be brought between allowed_lifetime_min and allowed_lifetime_max (inclusive). This means that, if the value of max_lifetime defined in the room's state is lower than allowed_lifetime_min, the value of allowed_lifetime_min will be used instead. Likewise, if the value of max_lifetime is higher than allowed_lifetime_max, the value of allowed_lifetime_max will be used instead.

In the example above, we ensure Synapse never deletes events that are less than one day old, and that it always deletes events that are over a year old.

If a default policy is set, and its max_lifetime value is lower than allowed_lifetime_min or higher than allowed_lifetime_max, the same process applies.

Both parameters are optional; if one is omitted Synapse won't use it to adjust the effective value of max_lifetime.

Like other settings in this section, these parameters can be expressed either as a duration or as a number of milliseconds.

Room configuration

To configure a room's message retention policy, a room's admin or moderator needs to send a state event in that room with the type m.room.retention and the following content:

{
    "max_lifetime": ...
}

In this event's content, the max_lifetime parameter has the same meaning as previously described, and needs to be expressed in milliseconds. The event's content can also include a min_lifetime parameter, which has the same meaning and limited support as previously described.

Note that over every server in the room, only the ones with support for message retention policies will actually remove expired events. This support is currently not enabled by default in Synapse.

Note on reclaiming disk space

While purge jobs actually delete data from the database, the disk space used by the database might not decrease immediately on the database's host. However, even though the database engine won't free up the disk space, it will start writing new data into where the purged data was.

If you want to reclaim the freed disk space anyway and return it to the operating system, the server admin needs to run VACUUM FULL; (or VACUUM; for SQLite databases) on Synapse's database (see the related PostgreSQL documentation).

Modules

Synapse supports extending its functionality by configuring external modules.

Note: When using third-party modules, you effectively allow someone else to run custom code on your Synapse homeserver. Server admins are encouraged to verify the provenance of the modules they use on their homeserver and make sure the modules aren't running malicious code on their instance.

Using modules

To use a module on Synapse, add it to the modules section of the configuration file:

modules:
  - module: my_super_module.MySuperClass
    config:
      do_thing: true
  - module: my_other_super_module.SomeClass
    config: {}

Each module is defined by a path to a Python class as well as a configuration. This information for a given module should be available in the module's own documentation.

Using multiple modules

The order in which modules are listed in this section is important. When processing an action that can be handled by several modules, Synapse will always prioritise the module that appears first (i.e. is the highest in the list). This means:

• If several modules register the same callback, the callback registered by the module that appears first is used.
• If several modules try to register a handler for the same HTTP path, only the handler registered by the module that appears first is used. Handlers registered by the other module(s) are ignored and Synapse will log a warning message about them.

Note that Synapse doesn't allow multiple modules implementing authentication checkers via the password auth provider feature for the same login type with different fields. If this happens, Synapse will refuse to start.

Current status

We are currently in the process of migrating module interfaces to this system. While some interfaces might be compatible with it, others still require configuring modules in another part of Synapse's configuration file.

Currently, only the following pre-existing interfaces are compatible with this new system:

• spam checker
• third-party rules
• presence router
• password auth providers

Writing a module

A module is a Python class that uses Synapse's module API to interact with the homeserver. It can register callbacks that Synapse will call on specific operations, as well as web resources to attach to Synapse's web server.

When instantiated, a module is given its parsed configuration as well as an instance of the synapse.module_api.ModuleApi class. The configuration is a dictionary, and is either the output of the module's parse_config static method (see below), or the configuration associated with the module in Synapse's configuration file.

See the documentation for the ModuleApi class here.

When Synapse runs with several modules configured

If Synapse is running with other modules configured, the order each module appears in within the modules section of the Synapse configuration file might restrict what it can or cannot register. See this section for more information.

On top of the rules listed in the link above, if a callback returns a value that should cause the current operation to fail (e.g. if a callback checking an event returns with a value that should cause the event to be denied), Synapse will fail the operation and ignore any subsequent callbacks that should have been run after this one.

The documentation for each callback mentions how Synapse behaves when multiple modules implement it.

Handling the module's configuration

A module can implement the following static method:

@staticmethod
def parse_config(config: dict) -> Any

This method is given a dictionary resulting from parsing the YAML configuration for the module. It may modify it (for example by parsing durations expressed as strings (e.g. "5d") into milliseconds, etc.), and return the modified dictionary. It may also verify that the configuration is correct, and raise an instance of synapse.module_api.errors.ConfigError if not.

Registering a web resource

Modules can register web resources onto Synapse's web server using the following module API method:

def ModuleApi.register_web_resource(path: str, resource: IResource) -> None

The path is the full absolute path to register the resource at. For example, if you register a resource for the path /_synapse/client/my_super_module/say_hello, Synapse will serve it at http(s)://[HS_URL]/_synapse/client/my_super_module/say_hello. Note that Synapse does not allow registering resources for several sub-paths in the /_matrix namespace (such as anything under /_matrix/client for example). It is strongly recommended that modules register their web resources under the /_synapse/client namespace.

The provided resource is a Python class that implements Twisted's IResource interface (such as Resource).

Only one resource can be registered for a given path. If several modules attempt to register a resource for the same path, the module that appears first in Synapse's configuration file takes priority.

Modules must register their web resources in their __init__ method.

Registering a callback

Modules can use Synapse's module API to register callbacks. Callbacks are functions that Synapse will call when performing specific actions. Callbacks must be asynchronous (unless specified otherwise), and are split in categories. A single module may implement callbacks from multiple categories, and is under no obligation to implement all callbacks from the categories it registers callbacks for.

Modules can register callbacks using one of the module API's register_[...]_callbacks methods. The callback functions are passed to these methods as keyword arguments, with the callback name as the argument name and the function as its value. A register_[...]_callbacks method exists for each category.

Callbacks for each category can be found on their respective page of the Synapse documentation website.

Spam checker callbacks

Spam checker callbacks allow module developers to implement spam mitigation actions for Synapse instances. Spam checker callbacks can be registered using the module API's register_spam_checker_callbacks method.

Callbacks

The available spam checker callbacks are:

check_event_for_spam

First introduced in Synapse v1.37.0

Changed in Synapse v1.60.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean or a string is now deprecated.

async def check_event_for_spam(event: "synapse.module_api.EventBase") -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", str, bool]

Called when receiving an event from a client or via federation. The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.
• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
• (deprecated) a non-Codes str to reject the operation and specify an error message. Note that clients typically will not localize the error message to the user's preferred locale.
• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

user_may_join_room

First introduced in Synapse v1.37.0

Changed in Synapse v1.61.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def user_may_join_room(user: str, room: str, is_invited: bool) -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when a user is trying to join a room. The user is represented by their Matrix user ID (e.g. @alice:example.com) and the room is represented by its Matrix ID (e.g. !room:example.com). The module is also given a boolean to indicate whether the user currently has a pending invite in the room.

This callback isn't called if the join is performed by a server administrator, or in the context of a room creation.

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.
• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

user_may_invite

First introduced in Synapse v1.37.0

Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def user_may_invite(inviter: str, invitee: str, room_id: str) -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when processing an invitation. Both inviter and invitee are represented by their Matrix user ID (e.g. @alice:example.com).

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.

• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.

• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.

• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

user_may_send_3pid_invite

First introduced in Synapse v1.45.0

Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def user_may_send_3pid_invite(
    inviter: str,
    medium: str,
    address: str,
    room_id: str,
) -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when processing an invitation using a third-party identifier (also called a 3PID, e.g. an email address or a phone number).

The inviter is represented by their Matrix user ID (e.g. @alice:example.com), and the invitee is represented by its medium (e.g. "email") and its address (e.g. alice@example.com). See the Matrix specification for more information regarding third-party identifiers.

For example, a call to this callback to send an invitation to the email address alice@example.com would look like this:

await user_may_send_3pid_invite(
    "@bob:example.com",  # The inviter's user ID
    "email",  # The medium of the 3PID to invite
    "alice@example.com",  # The address of the 3PID to invite
    "!some_room:example.com",  # The ID of the room to send the invite into
)

Note: If the third-party identifier is already associated with a matrix user ID, user_may_invite will be used instead.

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.

• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.

• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.

• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

user_may_create_room

First introduced in Synapse v1.37.0

Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def user_may_create_room(user_id: str) -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when processing a room creation request.

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.

• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.

• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.

• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

user_may_create_room_alias

First introduced in Synapse v1.37.0

Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def user_may_create_room_alias(user_id: str, room_alias: "synapse.module_api.RoomAlias") -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when trying to associate an alias with an existing room.

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.

• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.

• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.

• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

user_may_publish_room

First introduced in Synapse v1.37.0

Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def user_may_publish_room(user_id: str, room_id: str) -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when trying to publish a room to the homeserver's public rooms directory.

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.

• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.

• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.

• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

check_username_for_spam

First introduced in Synapse v1.37.0

async def check_username_for_spam(user_profile: synapse.module_api.UserProfile) -> bool

Called when computing search results in the user directory. The module must return a bool indicating whether the given user should be excluded from user directory searches. Return True to indicate that the user is spammy and exclude them from search results; otherwise return False.

The profile is represented as a dictionary with the following keys:

• user_id: str. The Matrix ID for this user.
• display_name: Optional[str]. The user's display name, or None if this user has not set a display name.
• avatar_url: Optional[str]. The mxc:// URL to the user's avatar, or None if this user has not set an avatar.

The module is given a copy of the original dictionary, so modifying it from within the module cannot modify a user's profile when included in user directory search results.

If multiple modules implement this callback, they will be considered in order. If a callback returns False, Synapse falls through to the next one. The value of the first callback that does not return False will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

check_registration_for_spam

First introduced in Synapse v1.37.0

async def check_registration_for_spam(
    email_threepid: Optional[dict],
    username: Optional[str],
    request_info: Collection[Tuple[str, str]],
    auth_provider_id: Optional[str] = None,
) -> "synapse.spam_checker_api.RegistrationBehaviour"

Called when registering a new user. The module must return a RegistrationBehaviour indicating whether the registration can go through or must be denied, or whether the user may be allowed to register but will be shadow banned.

The arguments passed to this callback are:

• email_threepid: The email address used for registering, if any.
• username: The username the user would like to register. Can be None, meaning that Synapse will generate one later.
• request_info: A collection of tuples, which first item is a user agent, and which second item is an IP address. These user agents and IP addresses are the ones that were used during the registration process.
• auth_provider_id: The identifier of the SSO authentication provider, if any.

If multiple modules implement this callback, they will be considered in order. If a callback returns RegistrationBehaviour.ALLOW, Synapse falls through to the next one. The value of the first callback that does not return RegistrationBehaviour.ALLOW will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

check_media_file_for_spam

First introduced in Synapse v1.37.0

Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.

async def check_media_file_for_spam(
    file_wrapper: "synapse.rest.media.v1.media_storage.ReadableFileWrapper",
    file_info: "synapse.rest.media.v1._base.FileInfo",
) -> Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes", bool]

Called when storing a local or remote file.

The callback must return one of:

• synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still decide to reject it.

• synapse.module_api.errors.Codes to reject the operation with an error code. In case of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.

• (deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.

• (deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.

If multiple modules implement this callback, they will be considered in order. If a callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one. The value of the first callback that does not return synapse.module_api.NOT_SPAM will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

should_drop_federated_event

First introduced in Synapse v1.60.0

async def should_drop_federated_event(event: "synapse.events.EventBase") -> bool

Called when checking whether a remote server can federate an event with us. Returning True from this function will silently drop a federated event and split-brain our view of a room's DAG, and thus you shouldn't use this callback unless you know what you are doing.

If multiple modules implement this callback, they will be considered in order. If a callback returns False, Synapse falls through to the next one. The value of the first callback that does not return False will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

Example

The example below is a module that implements the spam checker callback check_event_for_spam to deny any message sent by users whose Matrix user IDs are mentioned in a configured list, and registers a web resource to the path /_synapse/client/list_spam_checker/is_evil that returns a JSON object indicating whether the provided user appears in that list.

import json
from typing import Union

from twisted.web.resource import Resource
from twisted.web.server import Request

from synapse.module_api import ModuleApi


class IsUserEvilResource(Resource):
    def __init__(self, config):
        super(IsUserEvilResource, self).__init__()
        self.evil_users = config.get("evil_users") or []

    def render_GET(self, request: Request):
        user = request.args.get(b"user")[0].decode()
        request.setHeader(b"Content-Type", b"application/json")
        return json.dumps({"evil": user in self.evil_users}).encode()


class ListSpamChecker:
    def __init__(self, config: dict, api: ModuleApi):
        self.api = api
        self.evil_users = config.get("evil_users") or []

        self.api.register_spam_checker_callbacks(
            check_event_for_spam=self.check_event_for_spam,
        )

        self.api.register_web_resource(
            path="/_synapse/client/list_spam_checker/is_evil",
            resource=IsUserEvilResource(config),
        )

    async def check_event_for_spam(self, event: "synapse.events.EventBase") -> Union[Literal["NOT_SPAM"], Codes]:
        if event.sender in self.evil_users:
          return Codes.FORBIDDEN
        else:
          return synapse.module_api.NOT_SPAM

Third party rules callbacks

Third party rules callbacks allow module developers to add extra checks to verify the validity of incoming events. Third party event rules callbacks can be registered using the module API's register_third_party_rules_callbacks method.

Callbacks

The available third party rules callbacks are:

check_event_allowed

First introduced in Synapse v1.39.0

async def check_event_allowed(
    event: "synapse.events.EventBase",
    state_events: "synapse.types.StateMap",
) -> Tuple[bool, Optional[dict]]

This callback is very experimental and can and will break without notice. Module developers are encouraged to implement check_event_for_spam from the spam checker category instead.

Called when processing any incoming event, with the event and a StateMap representing the current state of the room the event is being sent into. A StateMap is a dictionary that maps tuples containing an event type and a state key to the corresponding state event. For example retrieving the room's m.room.create event from the state_events argument would look like this: state_events.get(("m.room.create", "")). The module must return a boolean indicating whether the event can be allowed.

Note that this callback function processes incoming events coming via federation traffic (on top of client traffic). This means denying an event might cause the local copy of the room's history to diverge from that of remote servers. This may cause federation issues in the room. It is strongly recommended to only deny events using this callback function if the sender is a local user, or in a private federation in which all servers are using the same module, with the same configuration.

If the boolean returned by the module is True, it may also tell Synapse to replace the event with new data by returning the new event's data as a dictionary. In order to do that, it is recommended the module calls event.get_dict() to get the current event as a dictionary, and modify the returned dictionary accordingly.

If check_event_allowed raises an exception, the module is assumed to have failed. The event will not be accepted but is not treated as explicitly rejected, either. An HTTP request causing the module check will likely result in a 500 Internal Server Error.

When the boolean returned by the module is False, the event is rejected. (Module developers should not use exceptions for rejection.)

Note that replacing the event only works for events sent by local users, not for events received over federation.

If multiple modules implement this callback, they will be considered in order. If a callback returns True, Synapse falls through to the next one. The value of the first callback that does not return True will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

on_create_room

First introduced in Synapse v1.39.0

async def on_create_room(
    requester: "synapse.types.Requester",
    request_content: dict,
    is_requester_admin: bool,
) -> None

Called when processing a room creation request, with the Requester object for the user performing the request, a dictionary representing the room creation request's JSON body (see the spec for a list of possible parameters), and a boolean indicating whether the user performing the request is a server admin.

Modules can modify the request_content (by e.g. adding events to its initial_state), or deny the room's creation by raising a module_api.errors.SynapseError.

If multiple modules implement this callback, they will be considered in order. If a callback returns without raising an exception, Synapse falls through to the next one. The room creation will be forbidden as soon as one of the callbacks raises an exception. If this happens, Synapse will not call any of the subsequent implementations of this callback.

check_threepid_can_be_invited

First introduced in Synapse v1.39.0

async def check_threepid_can_be_invited(
    medium: str,
    address: str,
    state_events: "synapse.types.StateMap",
) -> bool:

Called when processing an invite via a third-party identifier (i.e. email or phone number). The module must return a boolean indicating whether the invite can go through.

If multiple modules implement this callback, they will be considered in order. If a callback returns True, Synapse falls through to the next one. The value of the first callback that does not return True will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

check_visibility_can_be_modified

First introduced in Synapse v1.39.0

async def check_visibility_can_be_modified(
    room_id: str,
    state_events: "synapse.types.StateMap",
    new_visibility: str,
) -> bool:

Called when changing the visibility of a room in the local public room directory. The visibility is a string that's either "public" or "private". The module must return a boolean indicating whether the change can go through.

If multiple modules implement this callback, they will be considered in order. If a callback returns True, Synapse falls through to the next one. The value of the first callback that does not return True will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

on_new_event

First introduced in Synapse v1.47.0

async def on_new_event(
    event: "synapse.events.EventBase",
    state_events: "synapse.types.StateMap",
) -> None:

Called after sending an event into a room. The module is passed the event, as well as the state of the room after the event. This means that if the event is a state event, it will be included in this state.

Note that this callback is called when the event has already been processed and stored into the room, which means this callback cannot be used to deny persisting the event. To deny an incoming event, see check_event_for_spam instead.

If multiple modules implement this callback, Synapse runs them all in order.

check_can_shutdown_room

First introduced in Synapse v1.55.0

async def check_can_shutdown_room(
    user_id: str, room_id: str,
) -> bool:

Called when an admin user requests the shutdown of a room. The module must return a boolean indicating whether the shutdown can go through. If the callback returns False, the shutdown will not proceed and the caller will see a M_FORBIDDEN error.

If multiple modules implement this callback, they will be considered in order. If a callback returns True, Synapse falls through to the next one. The value of the first callback that does not return True will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

check_can_deactivate_user

First introduced in Synapse v1.55.0

async def check_can_deactivate_user(
    user_id: str, by_admin: bool,
) -> bool:

Called when the deactivation of a user is requested. User deactivation can be performed by an admin or the user themselves, so developers are encouraged to check the requester when implementing this callback. The module must return a boolean indicating whether the deactivation can go through. If the callback returns False, the deactivation will not proceed and the caller will see a M_FORBIDDEN error.

The module is passed two parameters, user_id which is the ID of the user being deactivated, and by_admin which is True if the request is made by a serve admin, and False otherwise.

If multiple modules implement this callback, they will be considered in order. If a callback returns True, Synapse falls through to the next one. The value of the first callback that does not return True will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

on_profile_update

First introduced in Synapse v1.54.0

async def on_profile_update(
    user_id: str,
    new_profile: "synapse.module_api.ProfileInfo",
    by_admin: bool,
    deactivation: bool,
) -> None:

Called after updating a local user's profile. The update can be triggered either by the user themselves or a server admin. The update can also be triggered by a user being deactivated (in which case their display name is set to an empty string ("") and the avatar URL is set to None). The module is passed the Matrix ID of the user whose profile has been updated, their new profile, as well as a by_admin boolean that is True if the update was triggered by a server admin (and False otherwise), and a deactivated boolean that is True if the update is a result of the user being deactivated.

Note that the by_admin boolean is also True if the profile change happens as a result of the user logging in through Single Sign-On, or if a server admin updates their own profile.

Per-room profile changes do not trigger this callback to be called. Synapse administrators wishing this callback to be called on every profile change are encouraged to disable per-room profiles globally using the allow_per_room_profiles configuration setting in Synapse's configuration file. This callback is not called when registering a user, even when setting it through the get_displayname_for_registration module callback.

If multiple modules implement this callback, Synapse runs them all in order.

on_user_deactivation_status_changed

First introduced in Synapse v1.54.0

async def on_user_deactivation_status_changed(
    user_id: str, deactivated: bool, by_admin: bool
) -> None:

Called after deactivating a local user, or reactivating them through the admin API. The deactivation can be triggered either by the user themselves or a server admin. The module is passed the Matrix ID of the user whose status is changed, as well as a deactivated boolean that is True if the user is being deactivated and False if they're being reactivated, and a by_admin boolean that is True if the deactivation was triggered by a server admin (and False otherwise). This latter by_admin boolean is always True if the user is being reactivated, as this operation can only be performed through the admin API.

If multiple modules implement this callback, Synapse runs them all in order.

on_threepid_bind

First introduced in Synapse v1.56.0

async def on_threepid_bind(user_id: str, medium: str, address: str) -> None:

Called after creating an association between a local user and a third-party identifier (email address, phone number). The module is given the Matrix ID of the user the association is for, as well as the medium (email or msisdn) and address of the third-party identifier.

Note that this callback is not called after a successful association on an identity server.

If multiple modules implement this callback, Synapse runs them all in order.

Example

The example below is a module that implements the third-party rules callback check_event_allowed to censor incoming messages as dictated by a third-party service.

from typing import Optional, Tuple

from synapse.module_api import ModuleApi

_DEFAULT_CENSOR_ENDPOINT = "https://my-internal-service.local/censor-event"

class EventCensorer:
    def __init__(self, config: dict, api: ModuleApi):
        self.api = api
        self._endpoint = config.get("endpoint", _DEFAULT_CENSOR_ENDPOINT)

        self.api.register_third_party_rules_callbacks(
            check_event_allowed=self.check_event_allowed,
        )

    async def check_event_allowed(
        self,
        event: "synapse.events.EventBase",
        state_events: "synapse.types.StateMap",
    ) -> Tuple[bool, Optional[dict]]:
        event_dict = event.get_dict()
        new_event_content = await self.api.http_client.post_json_get_json(
            uri=self._endpoint, post_json=event_dict,
        )
        event_dict["content"] = new_event_content
        return event_dict

Presence router callbacks

Presence router callbacks allow module developers to specify additional users (local or remote) to receive certain presence updates from local users. Presence router callbacks can be registered using the module API's register_presence_router_callbacks method.

Callbacks

The available presence router callbacks are:

get_users_for_states

First introduced in Synapse v1.42.0

async def get_users_for_states(
    state_updates: Iterable["synapse.api.UserPresenceState"],
) -> Dict[str, Set["synapse.api.UserPresenceState"]]

Requires get_interested_users to also be registered

Called when processing updates to the presence state of one or more users. This callback can be used to instruct the server to forward that presence state to specific users. The module must return a dictionary that maps from Matrix user IDs (which can be local or remote) to the UserPresenceState changes that they should be forwarded.

Synapse will then attempt to send the specified presence updates to each user when possible.

If multiple modules implement this callback, Synapse merges all the dictionaries returned by the callbacks. If multiple callbacks return a dictionary containing the same key, Synapse concatenates the sets associated with this key from each dictionary.

get_interested_users

First introduced in Synapse v1.42.0

async def get_interested_users(
    user_id: str
) -> Union[Set[str], "synapse.module_api.PRESENCE_ALL_USERS"]

Requires get_users_for_states to also be registered

Called when determining which users someone should be able to see the presence state of. This callback should return complementary results to get_users_for_state or the presence information may not be properly forwarded.

The callback is given the Matrix user ID for a local user that is requesting presence data and should return the Matrix user IDs of the users whose presence state they are allowed to query. The returned users can be local or remote.

Alternatively the callback can return synapse.module_api.PRESENCE_ALL_USERS to indicate that the user should receive updates from all known users.

If multiple modules implement this callback, they will be considered in order. Synapse calls each callback one by one, and use a concatenation of all the sets returned by the callbacks. If one callback returns synapse.module_api.PRESENCE_ALL_USERS, Synapse uses this value instead. If this happens, Synapse does not call any of the subsequent implementations of this callback.

Example

The example below is a module that implements both presence router callbacks, and ensures that @alice:example.org receives all presence updates from @bob:example.com and @charlie:somewhere.org, regardless of whether Alice shares a room with any of them.

from typing import Dict, Iterable, Set, Union

from synapse.module_api import ModuleApi


class CustomPresenceRouter:
    def __init__(self, config: dict, api: ModuleApi):
        self.api = api

        self.api.register_presence_router_callbacks(
            get_users_for_states=self.get_users_for_states,
            get_interested_users=self.get_interested_users,
        )

    async def get_users_for_states(
        self,
        state_updates: Iterable["synapse.api.UserPresenceState"],
    ) -> Dict[str, Set["synapse.api.UserPresenceState"]]:
        res = {}
        for update in state_updates:
            if (
                update.user_id == "@bob:example.com"
                or update.user_id == "@charlie:somewhere.org"
            ):
                res.setdefault("@alice:example.com", set()).add(update)

        return res

    async def get_interested_users(
        self,
        user_id: str,
    ) -> Union[Set[str], "synapse.module_api.PRESENCE_ALL_USERS"]:
        if user_id == "@alice:example.com":
            return {"@bob:example.com", "@charlie:somewhere.org"}

        return set()

Account validity callbacks

Account validity callbacks allow module developers to add extra steps to verify the validity on an account, i.e. see if a user can be granted access to their account on the Synapse instance. Account validity callbacks can be registered using the module API's register_account_validity_callbacks method.

The available account validity callbacks are:

is_user_expired

First introduced in Synapse v1.39.0

async def is_user_expired(user: str) -> Optional[bool]

Called when processing any authenticated request (except for logout requests). The module can return a bool to indicate whether the user has expired and should be locked out of their account, or None if the module wasn't able to figure it out. The user is represented by their Matrix user ID (e.g. @alice:example.com).

If the module returns True, the current request will be denied with the error code ORG_MATRIX_EXPIRED_ACCOUNT and the HTTP status code 403. Note that this doesn't invalidate the user's access token.

If multiple modules implement this callback, they will be considered in order. If a callback returns None, Synapse falls through to the next one. The value of the first callback that does not return None will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

on_user_registration

First introduced in Synapse v1.39.0

async def on_user_registration(user: str) -> None

Called after successfully registering a user, in case the module needs to perform extra operations to keep track of them. (e.g. add them to a database table). The user is represented by their Matrix user ID.

If multiple modules implement this callback, Synapse runs them all in order.

Password auth provider callbacks

Password auth providers offer a way for server administrators to integrate their Synapse installation with an external authentication system. The callbacks can be registered by using the Module API's register_password_auth_provider_callbacks method.

Callbacks

auth_checkers

First introduced in Synapse v1.46.0

auth_checkers: Dict[Tuple[str, Tuple[str, ...]], Callable]

A dict mapping from tuples of a login type identifier (such as m.login.password) and a tuple of field names (such as ("password", "secret_thing")) to authentication checking callbacks, which should be of the following form:

async def check_auth(
    user: str,
    login_type: str,
    login_dict: "synapse.module_api.JsonDict",
) -> Optional[
    Tuple[
        str, 
        Optional[Callable[["synapse.module_api.LoginResponse"], Awaitable[None]]]
    ]
]

The login type and field names should be provided by the user in the request to the /login API. The Matrix specification defines some types, however user defined ones are also allowed.

The callback is passed the user field provided by the client (which might not be in @username:server form), the login type, and a dictionary of login secrets passed by the client.

If the authentication is successful, the module must return the user's Matrix ID (e.g. @alice:example.com) and optionally a callback to be called with the response to the /login request. If the module doesn't wish to return a callback, it must return None instead.

If the authentication is unsuccessful, the module must return None.

If multiple modules register an auth checker for the same login type but with different fields, Synapse will refuse to start.

If multiple modules register an auth checker for the same login type with the same fields, then the callbacks will be executed in order, until one returns a Matrix User ID (and optionally a callback). In that case, the return value of that callback will be accepted and subsequent callbacks will not be fired. If every callback returns None, then the authentication fails.

check_3pid_auth

First introduced in Synapse v1.46.0

async def check_3pid_auth(
    medium: str, 
    address: str,
    password: str,
) -> Optional[
    Tuple[
        str, 
        Optional[Callable[["synapse.module_api.LoginResponse"], Awaitable[None]]]
    ]
]

Called when a user attempts to register or log in with a third party identifier, such as email. It is passed the medium (eg. email), an address (eg. jdoe@example.com) and the user's password.

If the authentication is successful, the module must return the user's Matrix ID (e.g. @alice:example.com) and optionally a callback to be called with the response to the /login request. If the module doesn't wish to return a callback, it must return None instead.

If the authentication is unsuccessful, the module must return None.

If multiple modules implement this callback, they will be considered in order. If a callback returns None, Synapse falls through to the next one. The value of the first callback that does not return None will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback. If every callback returns None, the authentication is denied.

on_logged_out

First introduced in Synapse v1.46.0

async def on_logged_out(
    user_id: str,
    device_id: Optional[str],
    access_token: str
) -> None

Called during a logout request for a user. It is passed the qualified user ID, the ID of the deactivated device (if any: access tokens are occasionally created without an associated device ID), and the (now deactivated) access token.

If multiple modules implement this callback, Synapse runs them all in order.

get_username_for_registration

First introduced in Synapse v1.52.0

async def get_username_for_registration(
    uia_results: Dict[str, Any],
    params: Dict[str, Any],
) -> Optional[str]

Called when registering a new user. The module can return a username to set for the user being registered by returning it as a string, or None if it doesn't wish to force a username for this user. If a username is returned, it will be used as the local part of a user's full Matrix ID (e.g. it's alice in @alice:example.com).

This callback is called once User-Interactive Authentication has been completed by the user. It is not called when registering a user via SSO. It is passed two dictionaries, which include the information that the user has provided during the registration process.

The first dictionary contains the results of the User-Interactive Authentication flow followed by the user. Its keys are the identifiers of every step involved in the flow, associated with either a boolean value indicating whether the step was correctly completed, or additional information (e.g. email address, phone number...). A list of most existing identifiers can be found in the Matrix specification. Here's an example featuring all currently supported keys:

{
    "m.login.dummy": True,  # Dummy authentication
    "m.login.terms": True,  # User has accepted the terms of service for the homeserver
    "m.login.recaptcha": True,  # User has completed the recaptcha challenge
    "m.login.email.identity": {  # User has provided and verified an email address
        "medium": "email",
        "address": "alice@example.com",
        "validated_at": 1642701357084,
    },
    "m.login.msisdn": {  # User has provided and verified a phone number
        "medium": "msisdn",
        "address": "33123456789",
        "validated_at": 1642701357084,
    },
    "m.login.registration_token": "sometoken",  # User has registered through a registration token
}

The second dictionary contains the parameters provided by the user's client in the request to /_matrix/client/v3/register. See the Matrix specification for a complete list of these parameters.

If the module cannot, or does not wish to, generate a username for this user, it must return None.

If multiple modules implement this callback, they will be considered in order. If a callback returns None, Synapse falls through to the next one. The value of the first callback that does not return None will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback. If every callback returns None, the username provided by the user is used, if any (otherwise one is automatically generated).

get_displayname_for_registration

First introduced in Synapse v1.54.0

async def get_displayname_for_registration(
    uia_results: Dict[str, Any],
    params: Dict[str, Any],
) -> Optional[str]

Called when registering a new user. The module can return a display name to set for the user being registered by returning it as a string, or None if it doesn't wish to force a display name for this user.

This callback is called once User-Interactive Authentication has been completed by the user. It is not called when registering a user via SSO. It is passed two dictionaries, which include the information that the user has provided during the registration process. These dictionaries are identical to the ones passed to get_username_for_registration, so refer to the documentation of this callback for more information about them.

If multiple modules implement this callback, they will be considered in order. If a callback returns None, Synapse falls through to the next one. The value of the first callback that does not return None will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback. If every callback returns None, the username will be used (e.g. alice if the user being registered is @alice:example.com).

is_3pid_allowed

First introduced in Synapse v1.53.0

async def is_3pid_allowed(self, medium: str, address: str, registration: bool) -> bool

Called when attempting to bind a third-party identifier (i.e. an email address or a phone number). The module is given the medium of the third-party identifier (which is email if the identifier is an email address, or msisdn if the identifier is a phone number) and its address, as well as a boolean indicating whether the attempt to bind is happening as part of registering a new user. The module must return a boolean indicating whether the identifier can be allowed to be bound to an account on the local homeserver.

If multiple modules implement this callback, they will be considered in order. If a callback returns True, Synapse falls through to the next one. The value of the first callback that does not return True will be used. If this happens, Synapse will not call any of the subsequent implementations of this callback.

Example

The example module below implements authentication checkers for two different login types:

• my.login.type
  • Expects a my_field field to be sent to /login
  • Is checked by the method: self.check_my_login
• m.login.password (defined in the spec)
  • Expects a password field to be sent to /login
  • Is checked by the method: self.check_pass

from typing import Awaitable, Callable, Optional, Tuple

import synapse
from synapse import module_api


class MyAuthProvider:
    def __init__(self, config: dict, api: module_api):

        self.api = api

        self.credentials = {
            "bob": "building",
            "@scoop:matrix.org": "digging",
        }

        api.register_password_auth_provider_callbacks(
            auth_checkers={
                ("my.login_type", ("my_field",)): self.check_my_login,
                ("m.login.password", ("password",)): self.check_pass,
            },
        )

    async def check_my_login(
        self,
        username: str,
        login_type: str,
        login_dict: "synapse.module_api.JsonDict",
    ) -> Optional[
        Tuple[
            str,
            Optional[Callable[["synapse.module_api.LoginResponse"], Awaitable[None]]],
        ]
    ]:
        if login_type != "my.login_type":
            return None

        if self.credentials.get(username) == login_dict.get("my_field"):
            return (self.api.get_qualified_user_id(username), None)

    async def check_pass(
        self,
        username: str,
        login_type: str,
        login_dict: "synapse.module_api.JsonDict",
    ) -> Optional[
        Tuple[
            str,
            Optional[Callable[["synapse.module_api.LoginResponse"], Awaitable[None]]],
        ]
    ]:
        if login_type != "m.login.password":
            return None

        if self.credentials.get(username) == login_dict.get("password"):
            return (self.api.get_qualified_user_id(username), None)

Background update controller callbacks

Background update controller callbacks allow module developers to control (e.g. rate-limit) how database background updates are run. A database background update is an operation Synapse runs on its database in the background after it starts. It's usually used to run database operations that would take too long if they were run at the same time as schema updates (which are run on startup) and delay Synapse's startup too much: populating a table with a big amount of data, adding an index on a big table, deleting superfluous data, etc.

Background update controller callbacks can be registered using the module API's register_background_update_controller_callbacks method. Only the first module (in order of appearance in Synapse's configuration file) calling this method can register background update controller callbacks, subsequent calls are ignored.

The available background update controller callbacks are:

on_update

First introduced in Synapse v1.49.0

def on_update(update_name: str, database_name: str, one_shot: bool) -> AsyncContextManager[int]

Called when about to do an iteration of a background update. The module is given the name of the update, the name of the database, and a flag to indicate whether the background update will happen in one go and may take a long time (e.g. creating indices). If this last argument is set to False, the update will be run in batches.

The module must return an async context manager. It will be entered before Synapse runs a background update; this should return the desired duration of the iteration, in milliseconds.

The context manager will be exited when the iteration completes. Note that the duration returned by the context manager is a target, and an iteration may take substantially longer or shorter. If the one_shot flag is set to True, the duration returned is ignored.

Note: Unlike most module callbacks in Synapse, this one is synchronous. This is because asynchronous operations are expected to be run by the async context manager.

This callback is required when registering any other background update controller callback.

default_batch_size

First introduced in Synapse v1.49.0

async def default_batch_size(update_name: str, database_name: str) -> int

Called before the first iteration of a background update, with the name of the update and of the database. The module must return the number of elements to process in this first iteration.

If this callback is not defined, Synapse will use a default value of 100.

min_batch_size

First introduced in Synapse v1.49.0

async def min_batch_size(update_name: str, database_name: str) -> int

Called before running a new batch for a background update, with the name of the update and of the database. The module must return an integer representing the minimum number of elements to process in this iteration. This number must be at least 1, and is used to ensure that progress is always made.

If this callback is not defined, Synapse will use a default value of 100.

Account data callbacks

Account data callbacks allow module developers to react to changes of the account data of local users. Account data callbacks can be registered using the module API's register_account_data_callbacks method.

Callbacks

The available account data callbacks are:

on_account_data_updated

First introduced in Synapse v1.57.0

async def on_account_data_updated(
    user_id: str,
    room_id: Optional[str],
    account_data_type: str,
    content: "synapse.module_api.JsonDict",
) -> None:

Called after user's account data has been updated. The module is given the Matrix ID of the user whose account data is changing, the room ID the data is associated with, the type associated with the change, as well as the new content. If the account data is not associated with a specific room, then the room ID is None.

This callback is triggered when new account data is added or when the data associated with a given type (and optionally room) changes. This includes deletion, since in Matrix, deleting account data consists of replacing the data associated with a given type (and optionally room) with an empty dictionary ({}).

Note that this doesn't trigger when changing the tags associated with a room, as these are processed separately by Synapse.

If multiple modules implement this callback, Synapse runs them all in order.

Example

The example below is a module that implements the on_account_data_updated callback, and sends an event to an audit room when a user changes their account data.

import json
import attr
from typing import Any, Dict, Optional

from synapse.module_api import JsonDict, ModuleApi
from synapse.module_api.errors import ConfigError


@attr.s(auto_attribs=True)
class CustomAccountDataConfig:
    audit_room: str
    sender: str


class CustomAccountDataModule:
    def __init__(self, config: CustomAccountDataConfig, api: ModuleApi):
        self.api = api
        self.config = config

        self.api.register_account_data_callbacks(
            on_account_data_updated=self.log_new_account_data,
        )

    @staticmethod
    def parse_config(config: Dict[str, Any]) -> CustomAccountDataConfig:
        def check_in_config(param: str):
            if param not in config:
                raise ConfigError(f"'{param}' is required")

        check_in_config("audit_room")
        check_in_config("sender")

        return CustomAccountDataConfig(
            audit_room=config["audit_room"],
            sender=config["sender"],
        )

    async def log_new_account_data(
        self,
        user_id: str,
        room_id: Optional[str],
        account_data_type: str,
        content: JsonDict,
    ) -> None:
        content_raw = json.dumps(content)
        msg_content = f"{user_id} has changed their account data for type {account_data_type} to: {content_raw}"

        if room_id is not None:
            msg_content += f" (in room {room_id})"

        await self.api.create_and_send_event_into_room(
            {
                "room_id": self.config.audit_room,
                "sender": self.config.sender,
                "type": "m.room.message",
                "content": {
                    "msgtype": "m.text",
                    "body": msg_content
                }
            }
        )

Porting an existing module that uses the old interface

In order to port a module that uses Synapse's old module interface, its author needs to:

• ensure the module's callbacks are all asynchronous.
• register their callbacks using one or more of the register_[...]_callbacks methods from the ModuleApi class in the module's __init__ method (see this section for more info).

Additionally, if the module is packaged with an additional web resource, the module should register this resource in its __init__ method using the register_web_resource method from the ModuleApi class (see this section for more info).

There is no longer a get_db_schema_files callback provided for password auth provider modules. Any changes to the database should now be made by the module using the module API class.

The module's author should also update any example in the module's configuration to only use the new modules section in Synapse's configuration file (see this section for more info).

Scaling synapse via workers

For small instances it is recommended to run Synapse in the default monolith mode. For larger instances where performance is a concern it can be helpful to split out functionality into multiple separate python processes. These processes are called 'workers', and are (eventually) intended to scale horizontally independently.

Synapse's worker support is under active development and subject to change as we attempt to rapidly scale ever larger Synapse instances. However we are documenting it here to help admins needing a highly scalable Synapse instance similar to the one running matrix.org.

All processes continue to share the same database instance, and as such, workers only work with PostgreSQL-based Synapse deployments. SQLite should only be used for demo purposes and any admin considering workers should already be running PostgreSQL.

See also Matrix.org blog post for a higher level overview.

Main process/worker communication

The processes communicate with each other via a Synapse-specific protocol called 'replication' (analogous to MySQL- or Postgres-style database replication) which feeds streams of newly written data between processes so they can be kept in sync with the database state.

When configured to do so, Synapse uses a Redis pub/sub channel to send the replication stream between all configured Synapse processes. Additionally, processes may make HTTP requests to each other, primarily for operations which need to wait for a reply ─ such as sending an event.

All the workers and the main process connect to Redis, which relays replication commands between processes.

If Redis support is enabled Synapse will use it as a shared cache, as well as a pub/sub mechanism.

See the Architectural diagram section at the end for a visualisation of what this looks like.

Setting up workers

A Redis server is required to manage the communication between the processes. The Redis server should be installed following the normal procedure for your distribution (e.g. apt install redis-server on Debian). It is safe to use an existing Redis deployment if you have one.

Once installed, check that Redis is running and accessible from the host running Synapse, for example by executing echo PING | nc -q1 localhost 6379 and seeing a response of +PONG.

The appropriate dependencies must also be installed for Synapse. If using a virtualenv, these can be installed with:

pip install "matrix-synapse[redis]"

Note that these dependencies are included when synapse is installed with pip install matrix-synapse[all]. They are also included in the debian packages from matrix.org and in the docker images at https://hub.docker.com/r/matrixdotorg/synapse/.

To make effective use of the workers, you will need to configure an HTTP reverse-proxy such as nginx or haproxy, which will direct incoming requests to the correct worker, or to the main synapse instance. See the reverse proxy documentation for information on setting up a reverse proxy.

When using workers, each worker process has its own configuration file which contains settings specific to that worker, such as the HTTP listener that it provides (if any), logging configuration, etc.

Normally, the worker processes are configured to read from a shared configuration file as well as the worker-specific configuration files. This makes it easier to keep common configuration settings synchronised across all the processes.

The main process is somewhat special in this respect: it does not normally need its own configuration file and can take all of its configuration from the shared configuration file.

Shared configuration

Normally, only a couple of changes are needed to make an existing configuration file suitable for use with workers. First, you need to enable an "HTTP replication listener" for the main process; and secondly, you need to enable redis-based replication. Optionally, a shared secret can be used to authenticate HTTP traffic between workers. For example:

# extend the existing `listeners` section. This defines the ports that the
# main process will listen on.
listeners:
  # The HTTP replication port
  - port: 9093
    bind_address: '127.0.0.1'
    type: http
    resources:
     - names: [replication]

# Add a random shared secret to authenticate traffic.
worker_replication_secret: ""

redis:
    enabled: true

See the configuration manual for the full documentation of each option.

Under no circumstances should the replication listener be exposed to the public internet; replication traffic is:

• always unencrypted
• unauthenticated, unless worker_replication_secret is configured

Worker configuration

In the config file for each worker, you must specify:

• The type of worker (worker_app). The currently available worker applications are listed below.
• A unique name for the worker (worker_name).
• The HTTP replication endpoint that it should talk to on the main synapse process (worker_replication_host and worker_replication_http_port)
• If handling HTTP requests, a worker_listeners option with an http listener, in the same way as the listeners option in the shared config.
• If handling the ^/_matrix/client/v3/keys/upload endpoint, the HTTP URI for the main process (worker_main_http_uri).

For example:

worker_app: synapse.app.generic_worker
worker_name: generic_worker1

# The replication listener on the main synapse process.
worker_replication_host: 127.0.0.1
worker_replication_http_port: 9093

worker_main_http_uri: http://localhost:8008/

worker_listeners:
  - type: http
    port: 8083
    resources:
      - names: [client, federation]

worker_log_config: /etc/matrix-synapse/generic-worker-log.yaml

...is a full configuration for a generic worker instance, which will expose a plain HTTP endpoint on port 8083 separately serving various endpoints, e.g. /sync, which are listed below.

Obviously you should configure your reverse-proxy to route the relevant endpoints to the worker (localhost:8083 in the above example).

Running Synapse with workers

Finally, you need to start your worker processes. This can be done with either synctl or your distribution's preferred service manager such as systemd. We recommend the use of systemd where available: for information on setting up systemd to start synapse workers, see Systemd with Workers. To use synctl, see Using synctl with Workers.

Available worker applications

synapse.app.generic_worker

This worker can handle API requests matching the following regular expressions. These endpoints can be routed to any worker. If a worker is set up to handle a stream then, for maximum efficiency, additional endpoints should be routed to that worker: refer to the stream writers section below for further information.

# Sync requests
^/_matrix/client/(r0|v3)/sync$
^/_matrix/client/(api/v1|r0|v3)/events$
^/_matrix/client/(api/v1|r0|v3)/initialSync$
^/_matrix/client/(api/v1|r0|v3)/rooms/[^/]+/initialSync$

# Federation requests
^/_matrix/federation/v1/event/
^/_matrix/federation/v1/state/
^/_matrix/federation/v1/state_ids/
^/_matrix/federation/v1/backfill/
^/_matrix/federation/v1/get_missing_events/
^/_matrix/federation/v1/publicRooms
^/_matrix/federation/v1/query/
^/_matrix/federation/v1/make_join/
^/_matrix/federation/v1/make_leave/
^/_matrix/federation/(v1|v2)/send_join/
^/_matrix/federation/(v1|v2)/send_leave/
^/_matrix/federation/(v1|v2)/invite/
^/_matrix/federation/v1/event_auth/
^/_matrix/federation/v1/exchange_third_party_invite/
^/_matrix/federation/v1/user/devices/
^/_matrix/key/v2/query
^/_matrix/federation/v1/hierarchy/

# Inbound federation transaction request
^/_matrix/federation/v1/send/

# Client API requests
^/_matrix/client/(api/v1|r0|v3|unstable)/createRoom$
^/_matrix/client/(api/v1|r0|v3|unstable)/publicRooms$
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/joined_members$
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/context/.*$
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/members$
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/state$
^/_matrix/client/v1/rooms/.*/hierarchy$
^/_matrix/client/unstable/org.matrix.msc2716/rooms/.*/batch_send$
^/_matrix/client/unstable/im.nheko.summary/rooms/.*/summary$
^/_matrix/client/(r0|v3|unstable)/account/3pid$
^/_matrix/client/(r0|v3|unstable)/account/whoami$
^/_matrix/client/(r0|v3|unstable)/devices$
^/_matrix/client/versions$
^/_matrix/client/(api/v1|r0|v3|unstable)/voip/turnServer$
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/event/
^/_matrix/client/(api/v1|r0|v3|unstable)/joined_rooms$
^/_matrix/client/(api/v1|r0|v3|unstable)/search$

# Encryption requests
# Note that ^/_matrix/client/(r0|v3|unstable)/keys/upload/ requires `worker_main_http_uri`
^/_matrix/client/(r0|v3|unstable)/keys/query$
^/_matrix/client/(r0|v3|unstable)/keys/changes$
^/_matrix/client/(r0|v3|unstable)/keys/claim$
^/_matrix/client/(r0|v3|unstable)/room_keys/
^/_matrix/client/(r0|v3|unstable)/keys/upload/

# Registration/login requests
^/_matrix/client/(api/v1|r0|v3|unstable)/login$
^/_matrix/client/(r0|v3|unstable)/register$
^/_matrix/client/v1/register/m.login.registration_token/validity$

# Event sending requests
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/redact
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/send
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/state/
^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/(join|invite|leave|ban|unban|kick)$
^/_matrix/client/(api/v1|r0|v3|unstable)/join/
^/_matrix/client/(api/v1|r0|v3|unstable)/profile/

# Account data requests
^/_matrix/client/(r0|v3|unstable)/.*/tags
^/_matrix/client/(r0|v3|unstable)/.*/account_data

# Receipts requests
^/_matrix/client/(r0|v3|unstable)/rooms/.*/receipt
^/_matrix/client/(r0|v3|unstable)/rooms/.*/read_markers

# Presence requests
^/_matrix/client/(api/v1|r0|v3|unstable)/presence/

# User directory search requests
^/_matrix/client/(r0|v3|unstable)/user_directory/search$

Additionally, the following REST endpoints can be handled for GET requests:

^/_matrix/client/(api/v1|r0|v3|unstable)/pushrules/

Pagination requests can also be handled, but all requests for a given room must be routed to the same instance. Additionally, care must be taken to ensure that the purge history admin API is not used while pagination requests for the room are in flight:

^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/messages$

Additionally, the following endpoints should be included if Synapse is configured to use SSO (you only need to include the ones for whichever SSO provider you're using):

# for all SSO providers
^/_matrix/client/(api/v1|r0|v3|unstable)/login/sso/redirect
^/_synapse/client/pick_idp$
^/_synapse/client/pick_username
^/_synapse/client/new_user_consent$
^/_synapse/client/sso_register$

# OpenID Connect requests.
^/_synapse/client/oidc/callback$

# SAML requests.
^/_synapse/client/saml2/authn_response$

# CAS requests.
^/_matrix/client/(api/v1|r0|v3|unstable)/login/cas/ticket$

Ensure that all SSO logins go to a single process. For multiple workers not handling the SSO endpoints properly, see #7530 and #9427.

Note that a HTTP listener with client and federation resources must be configured in the worker_listeners option in the worker config.

Load balancing

It is possible to run multiple instances of this worker app, with incoming requests being load-balanced between them by the reverse-proxy. However, different endpoints have different characteristics and so admins may wish to run multiple groups of workers handling different endpoints so that load balancing can be done in different ways.

For /sync and /initialSync requests it will be more efficient if all requests from a particular user are routed to a single instance. Extracting a user ID from the access token or Authorization header is currently left as an exercise for the reader. Admins may additionally wish to separate out /sync requests that have a since query parameter from those that don't (and /initialSync), as requests that don't are known as "initial sync" that happens when a user logs in on a new device and can be very resource intensive, so isolating these requests will stop them from interfering with other users ongoing syncs.

Federation and client requests can be balanced via simple round robin.

The inbound federation transaction request ^/_matrix/federation/v1/send/ should be balanced by source IP so that transactions from the same remote server go to the same process.

Registration/login requests can be handled separately purely to help ensure that unexpected load doesn't affect new logins and sign ups.

Finally, event sending requests can be balanced by the room ID in the URI (or the full URI, or even just round robin), the room ID is the path component after /rooms/. If there is a large bridge connected that is sending or may send lots of events, then a dedicated set of workers can be provisioned to limit the effects of bursts of events from that bridge on events sent by normal users.

Stream writers

Additionally, the writing of specific streams (such as events) can be moved off of the main process to a particular worker.

To enable this, the worker must have a HTTP replication listener configured, have a worker_name and be listed in the instance_map config. The same worker can handle multiple streams, but unless otherwise documented, each stream can only have a single writer.

For example, to move event persistence off to a dedicated worker, the shared configuration would include:

instance_map:
    event_persister1:
        host: localhost
        port: 8034

stream_writers:
    events: event_persister1

An example for a stream writer instance:

worker_app: synapse.app.generic_worker
worker_name: event_persister1 

# The replication listener on the main synapse process.
worker_replication_host: 127.0.0.1
worker_replication_http_port: 9093

worker_listeners:
  - type: http
    port: 8034
    resources:
      - names: [replication]

  # Enable listener if this stream writer handles endpoints for the `typing` or
  # `to_device` streams. Uses a different port to the `replication` listener to
  # avoid exposing the `replication` listener publicly.
  #
  #- type: http
  #  port: 8035
  #  resources:
  #    - names: [client]

worker_log_config: /etc/matrix-synapse/event-persister-log.yaml

Some of the streams have associated endpoints which, for maximum efficiency, should be routed to the workers handling that stream. See below for the currently supported streams and the endpoints associated with them:

The events stream

The events stream experimentally supports having multiple writers, where work is sharded between them by room ID. Note that you must restart all worker instances when adding or removing event persisters. An example stream_writers configuration with multiple writers:

stream_writers:
    events:
        - event_persister1
        - event_persister2

The typing stream

The following endpoints should be routed directly to the worker configured as the stream writer for the typing stream:

^/_matrix/client/(api/v1|r0|v3|unstable)/rooms/.*/typing

The to_device stream

The following endpoints should be routed directly to the worker configured as the stream writer for the to_device stream:

^/_matrix/client/(r0|v3|unstable)/sendToDevice/

The account_data stream

The following endpoints should be routed directly to the worker configured as the stream writer for the account_data stream:

^/_matrix/client/(r0|v3|unstable)/.*/tags
^/_matrix/client/(r0|v3|unstable)/.*/account_data

The receipts stream

The following endpoints should be routed directly to the worker configured as the stream writer for the receipts stream:

^/_matrix/client/(r0|v3|unstable)/rooms/.*/receipt
^/_matrix/client/(r0|v3|unstable)/rooms/.*/read_markers

The presence stream

The following endpoints should be routed directly to the worker configured as the stream writer for the presence stream:

^/_matrix/client/(api/v1|r0|v3|unstable)/presence/

Background tasks

There is also support for moving background tasks to a separate worker. Background tasks are run periodically or started via replication. Exactly which tasks are configured to run depends on your Synapse configuration (e.g. if stats is enabled).

To enable this, the worker must have a worker_name and can be configured to run background tasks. For example, to move background tasks to a dedicated worker, the shared configuration would include:

run_background_tasks_on: background_worker

You might also wish to investigate the update_user_directory_from_worker and media_instance_running_background_jobs settings.

An example for a dedicated background worker instance:

worker_app: synapse.app.generic_worker
worker_name: background_worker

# The replication listener on the main synapse process.
worker_replication_host: 127.0.0.1
worker_replication_http_port: 9093

worker_log_config: /etc/matrix-synapse/background-worker-log.yaml

Updating the User Directory

You can designate one generic worker to update the user directory.

Specify its name in the shared configuration as follows:

update_user_directory_from_worker: worker_name

This work cannot be load-balanced; please ensure the main process is restarted after setting this option in the shared configuration!

User directory updates allow REST endpoints matching the following regular expressions to work:

^/_matrix/client/(r0|v3|unstable)/user_directory/search$

The above endpoints can be routed to any worker, though you may choose to route it to the chosen user directory worker.

This style of configuration supersedes the legacy synapse.app.user_dir worker application type.

Notifying Application Services

You can designate one generic worker to send output traffic to Application Services.

Specify its name in the shared configuration as follows:

notify_appservices_from_worker: worker_name

This work cannot be load-balanced; please ensure the main process is restarted after setting this option in the shared configuration!

This style of configuration supersedes the legacy synapse.app.appservice worker application type.

synapse.app.pusher

Handles sending push notifications to sygnal and email. Doesn't handle any REST endpoints itself, but you should set start_pushers: False in the shared configuration file to stop the main synapse sending push notifications.

To run multiple instances at once the pusher_instances option should list all pusher instances by their worker name, e.g.:

pusher_instances:
    - pusher_worker1
    - pusher_worker2

synapse.app.appservice

Deprecated as of Synapse v1.59. Use synapse.app.generic_worker with the notify_appservices_from_worker option instead.

Handles sending output traffic to Application Services. Doesn't handle any REST endpoints itself, but you should set notify_appservices: False in the shared configuration file to stop the main synapse sending appservice notifications.

Note this worker cannot be load-balanced: only one instance should be active.

synapse.app.federation_sender

Handles sending federation traffic to other servers. Doesn't handle any REST endpoints itself, but you should set send_federation: False in the shared configuration file to stop the main synapse sending this traffic.

If running multiple federation senders then you must list each instance in the federation_sender_instances option by their worker_name. All instances must be stopped and started when adding or removing instances. For example:

federation_sender_instances:
    - federation_sender1
    - federation_sender2

synapse.app.media_repository

Handles the media repository. It can handle all endpoints starting with:

/_matrix/media/

... and the following regular expressions matching media-specific administration APIs:

^/_synapse/admin/v1/purge_media_cache$
^/_synapse/admin/v1/room/.*/media.*$
^/_synapse/admin/v1/user/.*/media.*$
^/_synapse/admin/v1/media/.*$
^/_synapse/admin/v1/quarantine_media/.*$
^/_synapse/admin/v1/users/.*/media$

You should also set enable_media_repo: False in the shared configuration file to stop the main synapse running background jobs related to managing the media repository. Note that doing so will prevent the main process from being able to handle the above endpoints.

In the media_repository worker configuration file, configure the http listener to expose the media resource. For example:

worker_listeners:
 - type: http
   port: 8085
   resources:
     - names:
       - media

Note that if running multiple media repositories they must be on the same server and you must configure a single instance to run the background tasks, e.g.:

media_instance_running_background_jobs: "media-repository-1"

Note that if a reverse proxy is used , then /_matrix/media/ must be routed for both inbound client and federation requests (if they are handled separately).

synapse.app.user_dir

Deprecated as of Synapse v1.59. Use synapse.app.generic_worker with the update_user_directory_from_worker option instead.

Handles searches in the user directory. It can handle REST endpoints matching the following regular expressions:

^/_matrix/client/(r0|v3|unstable)/user_directory/search$

When using this worker you must also set update_user_directory: false in the shared configuration file to stop the main synapse running background jobs related to updating the user directory.

Above endpoint is not required to be routed to this worker. By default, update_user_directory is set to true, which means the main process will handle updates. All workers configured with client can handle the above endpoint as long as either this worker or the main process are configured to handle it, and are online.

If update_user_directory is set to false, and this worker is not running, the above endpoint may give outdated results.

Historical apps

The following used to be separate worker application types, but are now equivalent to synapse.app.generic_worker:

• synapse.app.client_reader
• synapse.app.event_creator
• synapse.app.federation_reader
• synapse.app.frontend_proxy
• synapse.app.synchrotron

Migration from old config

A main change that has occurred is the merging of worker apps into synapse.app.generic_worker. This change is backwards compatible and so no changes to the config are required.

To migrate apps to use synapse.app.generic_worker simply update the worker_app option in the worker configs, and where worker are started (e.g. in systemd service files, but not required for synctl).

Architectural diagram