Retrieval activity endpoint hangs when source DB is unreachable
## Problem
The `/api/instance/retrieval` endpoint can hang indefinitely when the source database is unreachable or the restore container is stuck. This happens because `ReportActivity` is called with `context.Background()`, which has no timeout.
## Impact
- API calls to the retrieval activity endpoint can hang forever
- This can tie up HTTP handler goroutines and degrade DLE responsiveness
- Monitoring/health checks relying on this endpoint may stall
## Expected Behavior
The endpoint should return within a reasonable timeout (e.g., 15 seconds) even if the source database is unreachable.
## Fix
MR !1109 adds:
- A 15-second timeout context for the activity check
- A 10-second connect timeout for the pgx connection to the source database
issue