Improve observability of DSS instances #1257
Description
Is your feature request related to a problem? Please describe.
Feedback from real-life DSS deployments at least twice reported performance issues caused (possibly partially) by accumulation of stale data in the DB, leading to latency and contention resulting in degraded overall performances.
Those accumulations could probably have been detected before they had any performance impact. E.g. if the number of subscriptions was exposed as a metric and an alert was setup if it reaches a certain threshold.
Describe the solution you'd like
Solution would need to be discussed, but should cover:
- exposing some metrics (Prometheus, REST endpoint, in heartbeat, ...)
- alerting with some sane defaults
- operational documentation to configure those correctly, as values would probably change between deployments
Describe alternatives you've considered
N/A
Additional context
#1241
#1074
Discussed during InterUSS weekly 2025-08-26. Issue #1181 was mentioned as possibly relevant.