# Chat Switchboard — PrometheusRule alerts (v0.33.0)
# Source file for the Helm template. Deploy via:
#   monitoring.prometheusRule.enabled: true
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
  name: switchboard-alerts
spec:
  groups:
    - name: switchboard.rules
      rules:
        # Pod restart (possible OOM)
        - alert: SwitchboardPodRestart
          expr: increase(kube_pod_container_status_restarts_total{container="backend"}[1h]) > 0
          for: 0m
          labels:
            severity: warning
          annotations:
            summary: "Switchboard backend pod restarted (possible OOM)"
            description: "Container {{ $labels.container }} in pod {{ $labels.pod }} restarted."

        # Provider down for 5+ minutes
        - alert: SwitchboardProviderDown
          expr: switchboard_provider_status > 2
          for: 5m
          labels:
            severity: critical
          annotations:
            summary: "Provider {{ $labels.provider_config_id }} is down"

        # DB pool >80% utilized
        - alert: SwitchboardDBPoolExhaustion
          expr: switchboard_db_in_use_connections / switchboard_db_open_connections > 0.8
          for: 5m
          labels:
            severity: warning
          annotations:
            summary: "DB connection pool >80% utilized"

        # HTTP 5xx error rate >5%
        - alert: SwitchboardHighErrorRate
          expr: >
            sum(rate(switchboard_http_requests_total{status=~"5.."}[5m]))
            / sum(rate(switchboard_http_requests_total[5m])) > 0.05
          for: 5m
          labels:
            severity: warning
          annotations:
            summary: "HTTP 5xx error rate exceeds 5%"

        # Task failure rate >25%
        - alert: SwitchboardTaskFailureRate
          expr: >
            rate(switchboard_task_executions_total{status="error"}[15m])
            / rate(switchboard_task_executions_total[15m]) > 0.25
          for: 10m
          labels:
            severity: warning
          annotations:
            summary: "Task failure rate exceeds 25% over 15 minutes"

        # No completions processed in 15 minutes (canary)
        - alert: SwitchboardNoCompletions
          expr: sum(rate(switchboard_completions_total[10m])) == 0
          for: 15m
          labels:
            severity: critical
          annotations:
            summary: "No completions processed in 15 minutes"