Kubernetes Observability — Monitoring K8s Jobs
Kubernetes Observability — Tips — Monitoring K8s Job Workload Resources — Pod Lifecycle Recap
When monitoring the health of workloads in a Kubernetes Cluster we are often working with long-running server workloads where we are checking that Pods stay in the Running
Phase and that the Pod Status is Ready
.
However for effective alerts and visualisations we need to remember to take Kubernetes Jobs into account as they work a bit differently. Jobs will run to completion and then report the Failed
or Succeeded
Phase values.
Generally the condition to look further into a situation is if a Pod is not in the Ready
State AND not in the Succeeded
Phase for some period of time.
Corollarily make sure that a variety of Workload Resources are used for the testing of your Kubernetes Observability work to try to verify no unforeseen gaps or problems will occur.
Kubernetes Pod Lifecycle — Pod Phases
Kubernetes Jobs
Kubernetes Liveness and Readiness Probes
Workload Resources