kind practice matrix
Use kind for Release Radar practice. Keep Killercoda for polished exam-style scenarios; do not treat hosted labs as the version matrix source of truth.
Node images
Section titled “Node images”Pinned in docs/pins/kubernetes.yaml (adjust patch tags when you verify newer kindest publishes):
| Minor | kind node image | Notes |
|---|---|---|
| 1.35 | kindest/node:v1.35.8 | Matches current exam pin fixtures |
| 1.36 | kindest/node:v1.36.4 | Middle supported minor |
| 1.37 | kindest/node:v1.37.0 | Newest supported minor (Garhwal) |
Create separate clusters per minor (do not skew control plane vs workers beyond the version skew policy):
kind create cluster --name kd-135 --image kindest/node:v1.35.8kind create cluster --name kd-136 --image kindest/node:v1.36.4kind create cluster --name kd-137 --image kindest/node:v1.37.0Diff-only lab pattern
Section titled “Diff-only lab pattern”- Apply a baseline manifest on 1.35.
- Re-apply or patch on 1.36 / 1.37.
- Record the behavioral delta (API accepted, field default changed, removal error, new status).
- Do not re-teach “what is a Pod?”
Good first drills:
- 1.36: confirm
gitRepovolumes are rejected; migrate to initContainer +gitor git-sync. - 1.36: MutatingAdmissionPolicy / OCI image volumes — exercise GA paths from the 1.36 radar page.
- 1.37: explore KYAML output (
kubectlKYAML) and pod-level resource fields on a throwaway Deployment.
Cleanup
Section titled “Cleanup”kind delete cluster --name kd-135kind delete cluster --name kd-136kind delete cluster --name kd-137