maintenance checkpoint
A saved record of cluster state, such as node versions, readiness and budgets, taken before and after maintenance so the two can be compared.
In more detail
It turns did anything change into a diff instead of a memory. Record node versions and conditions, PodDisruptionBudgets and a smoke test result. If the after checkpoint differs unexpectedly, you know where to look.
Related terms
- drain — Cordoning a node and then evicting its pods through the eviction API so it can be safely maintained or removed.
- version skew — How far apart the versions of cluster components, such as the API server, kubelets and kubectl, are allowed to be.
- smoke test — A quick test that checks the most basic behaviour works after a deploy, before deeper testing or real traffic.
Taught in
- Overview Sprint: Four Findings Before The Auditor — Kubernetes and Cloud Native Security Associate (KCSA) Exam Prep
- Drain, Upgrade, Version Skew And Certificate Maintenance — Kubernetes: Production-Grade Container Orchestration