alerts-frontend.yaml

apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
  name: brukerdialog-frontend-alerts
  namespace: etterlatte
  labels:
    team: etterlatte
spec:
  groups:
    - name: alerts-frontend
      rules:
      - alert: Applikasjon nede
        expr: kube_deployment_status_replicas_available{deployment="\{{ APP_NAME }}"} == 0
        for: 2m
        annotations:
          consequence: "App {{ $labels.app }} er nede i namespace {{ $labels.kubernetes_namespace }}"
          action: "`kubectl describe pod -l app={{ $labels.app }} -n {{ $labels.namespace }}` for events, og `kubectl logs -l app={{ $labels.app }} -n {{ $labels.namespace }}` for logger"
        labels:
          namespace: etterlatte
          severity: critical

      - alert: Høy feilrate i logger
        expr: (100 * sum by (log_app, log_namespace) (rate(logd_messages_total{log_app="\{{ APP_NAME }}",log_level=~"Warning|Error"}[3m])) / sum by (log_app, log_namespace) (rate(logd_messages_total{log_app="\{{ APP_NAME }}"}[3m]))) > 10
        for: 3m
        annotations:
          consequence: "Høy feilrate i logger"
          action: "Sjekk loggene her: {{LOGS_URL}}"
        labels:
          namespace: etterlatte

      - alert: Feil i selftest # This alert uses a custom metric provided by https://github.com/navikt/common-java-modules
        expr: selftests_aggregate_result_status{app="\{{ APP_NAME }}"} > 0
        for: 1m
        annotations:
          consequence: "Feil i selftest"
          action: "Sjekk app {{ $labels.app }} i namespace {{ $labels.kubernetes_namespace }} sine selftest for å se hva som er galt"
        labels:
          namespace: etterlatte
          severity: critical