Upgrade hop
Kubernetes 1.31 → 1.32
1.32 is itself end of life. If you are on 1.31 today, plan the whole path to a supported version (at least 1.35) with the upgrade planner.
1 · Fix removed APIs first
2 kinds stop being served in 1.32. Update manifests, charts and controllers to the replacement versions while the cluster still runs 1.31: the replacements are already served there.
| Kind | API version | Stops being served | Use instead |
|---|---|---|---|
| FlowSchema | flowcontrol.apiserver.k8s.io/v1beta3 | 1.32 | flowcontrol.apiserver.k8s.io/v1 |
| PriorityLevelConfiguration | flowcontrol.apiserver.k8s.io/v1beta3 | 1.32 | flowcontrol.apiserver.k8s.io/v1 |
| ValidatingAdmissionPolicy | admissionregistration.k8s.io/v1alpha1 alpha · off |
1.32 | admissionregistration.k8s.io/v1 |
| ValidatingAdmissionPolicyBinding | admissionregistration.k8s.io/v1alpha1 alpha · off |
1.32 | admissionregistration.k8s.io/v1 |
| SelfSubjectReview | authentication.k8s.io/v1alpha1 alpha · off |
1.32 | authentication.k8s.io/v1 |
Field-level changes for each of these are on the API pages and in the official Deprecated API Migration Guide.
Check your files with the manifest checker; in a live cluster, the API server metric apiserver_requested_deprecated_apis shows which deprecated APIs clients still call.
2 · Other changes that can bite
Removed in 1.32
- flowcontrol.apiserver.k8s.io/v1beta3 (FlowSchema, PriorityLevelConfiguration) is no longer served; move manifests and clients to flowcontrol.apiserver.k8s.io/v1, available since v1.29.
- The original DRA implementation introduced in v1.26 is withdrawn; the structured-parameter model replaces it.
Our summary of the 1.32 release announcement (CC BY 4.0). Read the full changelog for your components.
3 · Upgrade in this order
- Prepare. Move every node to 1.29 or newer if any are older: a 1.32 API server supports kubelets back to 1.29 only. Upgrade to the newest 1.31 patch,
1.31.14. - kube-apiserver to 1.32 (in HA clusters, one instance at a time; old and new may differ by one minor during the roll). Use the newest patch,
1.32.13. - kube-controller-manager, kube-scheduler, cloud-controller-manager to 1.32, once every API server is on 1.32.
- kubelet on each node (drain first). Optional in this hop: kubelets from 1.29 to 1.32 all work with a 1.32 control plane.
- kube-proxy: no newer than the API server, and within the same window as the kubelet it runs beside.
- kubectl: 1.31 to 1.33 all talk to a 1.32 API server.
From the official version skew policy. Managed services run the control-plane steps for you; node pools are still yours to schedule.