feat(observability): add db query duration histogram and pool exhaustion monitoring

[winniebyte]

Summary

Implements two monitoring features:

Issue solutions-plug#964 - Database Query Duration Histogram

• Added db_query_duration_seconds Prometheus histogram with buckets [0.01, 0.05, 0.1, 0.25, 0.5, 1, 2.5, 5]
• Instrumented all key query paths in db.rs via with_timeout() to record query durations
• Added P50/P95/P99 query latency Grafana panel to SLO dashboard
• Added Prometheus alert DBSlowQueryP95 firing when P95 exceeds 500ms for 5 minutes

Issue solutions-plug#965 - Database Connection Pool Exhaustion Alert

• Added db_pool_exhaustion_total counter in metrics.rs, incremented on pool timeout errors
• Added Prometheus alert DBPoolExhaustion firing when rate exceeds 1/minute
• Added Grafana panel showing pool utilisation as a percentage
• Updated alerts.yaml with pool exhaustion alert rule

Testing

• Added unit tests for observe_db_query_duration and observe_db_pool_exhaustion metrics
• Existing db error tests remain passing

Closes solutions-plug#964
Closes solutions-plug#965