- Scheduler framework permit plugins now run at the end of the scheduling cycle, after reserve plugins. Waiting on permit will remain in the beginning of the binding cycle. ([#88199](https://github.com/kubernetes/kubernetes/pull/88199), [@mateuszlitwin](https://github.com/mateuszlitwin)) [SIG Scheduling]

* kubelet `--cert-dir` now defaults to `/var/lib/kubelet/pki`, in order to ensure bootstrapped and rotated certificates persist beyond a reboot. resolves an issue in kubeadm with false-positive `/var/lib/kubelet is not empty` message during pre-flight checks ([#53317](https://github.com/kubernetes/kubernetes/pull/53317), [@liggitt](https://github.com/liggitt))

* kubelet `--cert-dir` now defaults to `/var/lib/kubelet/pki`, in order to ensure bootstrapped and rotated certificates persist beyond a reboot. resolves an issue in kubeadm with false-positive `/var/lib/kubelet is not empty` message during pre-flight checks ([#53317](https://github.com/kubernetes/kubernetes/pull/53317), [@liggitt](https://github.com/liggitt))

## Changelog since v1.21.8

## Changes by Kind

### Feature

- Kube-apiserver: when merging lists, Server Side Apply now prefers the order of the submitted request instead of the existing persisted object ([#107569](https://github.com/kubernetes/kubernetes/pull/107569), [@jiahuif](https://github.com/jiahuif)) [SIG API Machinery, CLI, Cloud Provider, Cluster Lifecycle, Instrumentation, Storage and Testing]

### Bug or Regression

* fix typo  ([#59619](https://github.com/kubernetes/kubernetes/pull/59619), [@jianliao82](https://github.com/jianliao82))

* DaemonSet, Deployment, ReplicaSet, and StatefulSet objects are now persisted in etcd in apps/v1 format ([#58854](https://github.com/kubernetes/kubernetes/pull/58854), [@liggitt](https://github.com/liggitt))

  - pod eviction API calls returned spurious precondition errors and required a second evict API call to succeed
  - dry-run eviction API calls persisted a DisruptionTarget condition into the pod being evicted ([#116554](https://github.com/kubernetes/kubernetes/pull/116554), [@atiratree](https://github.com/atiratree))

- Add `plugin_execution_duration_seconds` metric for scheduler framework plugins. ([#84522](https://github.com/kubernetes/kubernetes/pull/84522), [@liu-cong](https://github.com/liu-cong))
- Add `permit_wait_duration_seconds` metric to the scheduler. ([#84011](https://github.com/kubernetes/kubernetes/pull/84011), [@liu-cong](https://github.com/liu-cong))

### Deprecated/changed metrics

- Add new `--server-dry-run` flag to `kubectl apply` so that the request will be sent to the server with the dry-run flag (alpha), which means that changes won't be persisted. ([#68069](https://github.com/kubernetes/kubernetes/pull/68069), [@apelisse](https://github.com/apelisse))

  - PriorityClass resources will no longer be served from `scheduling.k8s.io/v1beta1` and `scheduling.k8s.io/v1alpha1` in v1.17. Migrate to the `scheduling.k8s.io/v1` API, available since v1.14. Existing persisted data can be retrieved via the `scheduling.k8s.io/v1` API.

- Protobuf serialization of metav1.MicroTime timestamps (used in `Lease` and `Event` API objects) has been corrected to truncate to microsecond precision, to match the documented behavior and JSON/YAML serialization. Any existing persisted data is truncated to microsecond when read from etcd. ([#111936](https://github.com/kubernetes/kubernetes/pull/111936), [@haoruan](https://github.com/haoruan)) [SIG API Machinery]

### Feature