Managing the Kubernetes Cluster Lifecycle

Cluster upgrades have a way of turning into unplanned incidents. The version is three minor releases behind, the documentation is vague on ordering, and there's pressure to get it done before the maintenance window closes. The CKA distills this anxiety into a structured, timed task — which is actually a useful forcing function.

This article covers the core of cluster lifecycle management: upgrading a kubeadm cluster one minor version, including the version skew rules, the drain/cordon workflow, and the kubelet restart that kubeadm upgrade apply doesn't handle. For the two related lifecycle operations — etcd backup/restore and certificate management — see the dedicated lessons linked below.