System Health Dashboard
The System Health dashboard is admin-only and provides real-time visibility into platform health.
Health Components
- Database — connection pool stats (total, idle, waiting), latency
- Memory — heap usage percentage, RSS in MB, warning/critical thresholds
- API — total requests, error rate, latency P50/P95/P99
- AI — circuit breaker status (CLOSED/OPEN/HALF_OPEN), AI service connectivity
Anomaly Detection
The dashboard highlights anomalies:
- High latency on specific endpoints
- Rising error rates
- Memory pressure warnings
- Circuit breaker state changes