Add observability for DB operations

**Description**

```
Add lightweight metrics so you can spot slow queries and error spikes early:
timing per repo method or per query group
count failures by operation
optionally log slow queries above threshold

```

No files, tests, or entry points are named. Start by locating the database repository and query-operation entry points, then review any existing logging or metrics conventions; done should include timing, failure counts, and an agreed approach for optional slow-query logging.

Written by the indexing model from the issue text.