monitoring/alert_rules.yml

groups:
  - name: senti-alerts
    rules:
      - alert: SentiHighLatency
        expr: histogram_quantile(0.95, senti_query_latency_seconds_bucket) > 30
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "P95 latency > 30 seconds"

      - alert: SentiHighErrorRate
        expr: rate(http_requests_total{status=~"5.."}[5m]) > 0.05
        for: 2m
        labels:
          severity: critical
        annotations:
          summary: "Error rate > 5%"

      - alert: SentiSafetySpike
        expr: rate(senti_safety_events_total[5m]) > 1
        for: 1m
        labels:
          severity: warning
        annotations:
          summary: "Unusual safety event rate"