Skip to content
Prometheus Cheatsheet

Job / CronJob

Job completati

Section titled “Job completati”
kube_job_status_succeeded{namespace!~"kube-.*|openshift-.*"}

Job falliti

Section titled “Job falliti”
kube_job_status_failed{namespace!~"kube-.*|openshift-.*"}

Job falliti (alert-ready)

Section titled “Job falliti (alert-ready)”
kube_job_status_failed > 0

Job in completamento lento (completions attese non raggiunte)

Section titled “Job in completamento lento (completions attese non raggiunte)”
kube_job_spec_completions - kube_job_status_succeeded - kube_job_status_failed > 0

Pod short-lived (firma tipica dei job batch)

Section titled “Pod short-lived (firma tipica dei job batch)”
count_over_time(
  kube_pod_container_status_running{namespace!~"kube-.*|openshift-.*"}[15m]
) < 15

CronJob

Section titled “CronJob”

CronJob sospeso

Section titled “CronJob sospeso”
kube_cronjob_spec_suspend != 0

CronJob in ritardo (next schedule superato da >1h)

Section titled “CronJob in ritardo (next schedule superato da >1h)”
time() - kube_cronjob_next_schedule_time > 3600

Ultima esecuzione riuscita (timestamp)

Section titled “Ultima esecuzione riuscita (timestamp)”
kube_cronjob_status_last_successful_time{namespace!~"kube-.*|openshift-.*"}

CPU/Memoria per job (breakdown per container, utile se il job ha init-container o sidecar)

Section titled “CPU/Memoria per job (breakdown per container, utile se il job ha init-container o sidecar)”
sum by (namespace, pod, container) (
  rate(container_cpu_usage_seconds_total{
    namespace="<NS>", pod=~"<JOB_NAME>-.*",
    container!="", container!="POD"
  }[5m])
)