Skip to content

Observability & Health

A cluster you can't see into is a cluster you can only guess about. This section covers the layers Kubernetes gives you out of the box — health probes that decide whether a container is alive and ready, the logging and metrics pipelines built into every cluster, and the kubectl workflow for turning "something's wrong" into "here's exactly what's wrong" — before you ever reach for a dedicated monitoring stack.

Read in this order

  1. Probes: Liveness, Readiness, and Startup — what each probe type actually controls, and the misconfiguration that causes restart-loop death spirals
  2. Logging — how container output becomes something kubectl logs can show you, node-level rotation, and cluster-wide logging patterns
  3. Metrics and Metrics Server — installing metrics-server, kubectl top, and what the Resource Metrics API can and can't feed into autoscaling
  4. Events and Debugging — a repeatable methodology for going from a broken pod to a root cause, using tools built into every cluster

Build the habit before you need it

The debugging methodology in the last page of this section — status, describe, events, logs, exec, in that order — is worth memorizing before you're paged at 2am. Troubleshooting applies exactly this sequence to specific failure symptoms.

Next

With health, logs, and metrics covered, continue to Cluster Administration to see how these signals feed into running the cluster itself.