- Client-go: Optimized leaders renewing leases by updating leader lock optimistically without getting the record from the API server first. Also, a new metric `leader_election_slowpath_total` was added to allow users to monitor how many leader elections are updated non-optimistically.

   *
   * <p>Entries in the map can be in the following states:
   *
   * <p>Valid: - Live: valid key/value are set
   *
   * <p>Invalid: - Collected: key/value was partially collected, but not yet cleaned up
   */
  interface InternalEntry<K, V, E extends InternalEntry<K, V, E>> {
    /** Gets the next entry in the chain. */
    E getNext();

    /** Gets the entry's hash. */
    int getHash();

- Fix a bug in the dual-stack IPVS proxier where stale IPv6 endpoints were not being cleaned up ([#87695](https://github.com/kubernetes/kubernetes/pull/87695), [@andrewsykim](https://github.com/andrewsykim)) [SIG Network]

- Fix the regression with the slow pods termination. Before this fix pods may take an additional time to terminate - up to one minute. Reversing the change that ensured that CNI resources cleaned up when the pod is removed on API server. ([#97980](https://github.com/kubernetes/kubernetes/pull/97980), [@SergeyKanzhelev](https://github.com/SergeyKanzhelev)) [SIG Node]

* Fix metrics API group name in audit configuration ([#53493](https://github.com/kubernetes/kubernetes/pull/53493), [@piosz](https://github.com/piosz))
* Use separate client for leader election in scheduler to avoid starving leader election by regular scheduler operations. ([#53793](https://github.com/kubernetes/kubernetes/pull/53793), [@wojtek-t](https://github.com/wojtek-t))

	// and allowed to save.
	z.poolMeta = newPoolMeta(z, poolMeta{})
	z.poolMeta.dontSave = true

	bootstrapTrace("newSharedLock", func() {
		globalLeaderLock = newSharedLock(GlobalContext, z, "leader.lock")
	})

	// Start self healing after the object initialization
	// so various tasks will be useful
	bootstrapTrace("initAutoHeal", func() {
		initAutoHeal(GlobalContext, z)
	})

- Updated the default API priority-and-fairness config to avoid endpoint/configmaps operations from controller-manager to all match leader-election priority level. ([#106725](https://github.com/kubernetes/kubernetes/pull/106725), [@wojtek-t](https://github.com/wojtek-t))

* Allow disable outbound SNAT when Azure standard load balancer is used together with outbound rules.  ([#75282](https://github.com/kubernetes/kubernetes/pull/75282), [@feiskyer](https://github.com/feiskyer))
* Ensure Azure load balancer cleaned up on 404 or 403 when deleting LoadBalancer services. ([#75256](https://github.com/kubernetes/kubernetes/pull/75256), [@feiskyer](https://github.com/feiskyer))

- The `--leader-elect*` CLI args are now honored in scheduler. ([#105915](https://github.com/kubernetes/kubernetes/pull/105915), [@Huang-Wei](https://github.com/Huang-Wei))

  reclaimed.
  - This feature prevents the node and scheduler from oversubscribing resources
    that are still being used by a pod in the process of being cleaned up.
  - This feature can result in increase of Pod Deletion latency especially if the
    pod has a large filesystem footprint.

### RBAC