The back-end Solr cluster in production sometimes struggles to keep up with the current traffic. It would be helpful to have more insight at back-end Solr cluster level about search requests for profiling purposes. For example, the following metadata:

1. Endpoint: search page, taglookup page, ws/2, autocomplete, and medium/cdstub search for tracklist import in release editor.
2. Web Server: musicbrainz.org, beta.musicbrainz.org, and so on, as defined in DBDefs->WEB_SERVER.
3. Node: Docker host, optional, if defined in environment variable $MUSICBRAINZ_NODE_NAME.
4. Container: Docker container ID, optional, assuming host name in container.
5. Version: Git commit hash and Git branch.

Each of these should be conveyed as a different HTTP header, to make it easy for parsing through HAProxy and any log analysis tools.

There is no need to convey any client-sided metadata, such as IP address, user agent...