k8s.io/test-infra@v0.0.0-20240520184403-27c6b4c223d8/config/prow/cluster/monitoring/mixins/prometheus/ci_absent_alerts.libsonnet (about) 1 { 2 prometheusAlerts+:: { 3 groups+: [ 4 { 5 name: 'ci-absent', 6 rules: [ 7 { 8 alert: '%sDown' % name, 9 expr: ||| 10 absent(up{job="%s"} == 1) 11 ||| % name, 12 'for': '10m', 13 labels: { 14 severity: 'critical', 15 slo: name, 16 }, 17 annotations: { 18 message: '@test-infra-oncall The service %s has been down for 10 minutes.' % name, 19 }, 20 } 21 for name in $._config.ciAbsents.components 22 ], 23 }, 24 ], 25 }, 26 }