# Setting up Synapse with Workers using Docker Compose
This directory describes how deploy and manage Synapse and workers via [Docker Compose](https://docs.docker.com/compose/).
Example worker configuration files can be found [here](workers).
All examples and snippets assume that your Synapse service is called `synapse` in your Docker Compose file.
An example Docker Compose file can be found [here](docker-compose.yaml).
## Worker Service Examples in Docker Compose
In order to start the Synapse container as a worker, you must specify an `entrypoint` that loads both the `homeserver.yaml` and the configuration for the worker (`synapse-generic-worker-1.yaml` in the example below). You must also include the worker type in the environment variable `SYNAPSE_WORKER` or alternatively pass `-m synapse.app.generic_worker` as part of the `entrypoint` after `"/start.py", "run"`).
This assumes that your Redis service is called `redis` in your Docker Compose file.
### Add a replication Listener
Locate the `listeners` section of your `homeserver.yaml` and add the following replication listener:
```yaml
listeners:
# Other listeners
- port: 9093
type: http
resources:
- names: [replication]
```
This listener is used by the workers for replication and is referred to in worker config files using the following settings:
```yaml
worker_replication_host: synapse
worker_replication_http_port: 9093
```
### Configure Federation Senders
This section is applicable if you are using Federation senders (synapse.app.federation_sender). Locate the `send_federation` and `federation_sender_instances` settings in your `homeserver.yaml` and configure them:
```yaml
# This will disable federation sending on the main Synapse instance
send_federation: false
federation_sender_instances:
- synapse-federation-sender-1 # The worker_name setting in your federation sender worker configuration file
Using the concepts shown here it is possible to create other worker types in Docker Compose. See the [Workers](https://element-hq.github.io/synapse/latest/workers.html#available-worker-applications) documentation for a list of available workers.