[synapse-usage-exporter](https://github.com/loelkes/synapse-usage-exporter) allows you to export the usage statistics of a Synapse homeserver to this container service and for the collected metrics to later be scraped by Prometheus.
Synapse does not include usage statistics in its Prometheus metrics. They can be reported to an HTTP `PUT` endpoint 5 minutes after startup and from then on at a fixed interval of once every three hours. This role integrates a simple [Flask](https://flask.palletsprojects.com) project that offers an HTTP `PUT` endpoint and holds the most recent received record available to be scraped by Prometheus.
Enabling this service will automatically:
- install the synapse-usage-exporter service
- re-configure Synapse to push (via HTTP `PUT`) usage statistics information to synapse-usage-exporter
- re-configure [Prometheus](./configuring-playbook-prometheus-grafana.md) (if Prometheus is enabled), to periodically scrape metrics from synapse-usage-exporter
By default, this playbook installs synapse-usage-exporter on the `matrix.` subdomain, at the `/report-usage-stats/push` path (https://matrix.example.com/report-usage-stats/push). This makes it easy to install it, because it **doesn't require additional DNS records to be set up**. If that's okay, you can skip this section.
By tweaking the `matrix_synapse_usage_exporter_hostname` and `matrix_synapse_usage_exporter_path_prefix` variables, you can easily make the service available at a **different hostname and/or path** than the default one.
Example additional configuration for your `inventory/host_vars/matrix.example.com/vars.yml` file:
```yaml
# Change the default hostname and path prefix
# These variables are used only if (`matrix_synapse_usage_exporter_proxying_enabled: true`)
After configuring the playbook and potentially [adjusting your DNS records](#adjusting-dns-records), run the playbook with [playbook tags](playbook-tags.md) as below:
<!-- NOTE: let this conservative command run (instead of install-all) to make it clear that failure of the command means something is clearly broken. -->
`just install-all` is useful for maintaining your setup quickly ([2x-5x faster](../CHANGELOG.md#2x-5x-performance-improvements-in-playbook-runtime) than `just setup-all`) when its components remain unchanged. If you adjust your `vars.yml` to remove other components, you'd need to run `just setup-all`, or these components will still remain installed. Note these shortcuts run the `ensure-matrix-users-created` tag too.