Monitoring

Measure system health with metrics and alerts.

# Prometheus alert example
- alert: HighErrorRate
  expr: sum(rate(http_requests_total{status=~"5.."}[5m])) > 5
  for: 10m
  labels:
    severity: critical