feature request: add ability to expire old alert events based on some retention policy

It would be great to add the ability to expire old alert events based on some retention policy.

The issue does not name files, tests, or entry points. Start by locating how alert events are stored and queried, then determine where a retention policy could be configured and enforced. Done would require an agreed policy scope, configuration behavior, expiration mechanism, and tests covering old events.

Written by the indexing model from the issue text.