the garbage collector starts to manage it.


#### Monitoring/Prometheus

- Fix etcd_object_counts metric reported by kube-apiserver ([#94773](https://github.com/kubernetes/kubernetes/pull/94773), [@tkashem](https://github.com/tkashem)) [SIG API Machinery]
- Fix incorrectly reported verbs for kube-apiserver metrics for CRD objects ([#93523](https://github.com/kubernetes/kubernetes/pull/93523), [@wojtek-t](https://github.com/wojtek-t)) [SIG API Machinery and Instrumentation]

           rules:
           - apiGroups: [""]
             resources: ["events"]
             verbs:
             - create
             - patch
             - update
           - apiGroups: [""]
             resources: ["secrets"]
             verbs:
             - get
             - create
         - apiVersion: rbac.authorization.k8s.io/v1
           kind: ClusterRoleBinding

* The --long-running-request-regexp flag to kube-apiserver is deprecated and will be removed in a future release. Long-running requests are now detected based on specific verbs (watch, proxy) or subresources (proxy, portforward, log, exec, attach). ([#38119](https://github.com/kubernetes/kubernetes/pull/38119), [@liggitt](https://github.com/liggitt))

* Make sure that resources being updated are handled correctly by Quota system ([#52452](https://github.com/kubernetes/kubernetes/pull/52452), [@gnufied](https://github.com/gnufied))

* Fix panic for attaching AzureDisk to vmss nodes ([#63275](https://github.com/kubernetes/kubernetes/pull/63275), [@feiskyer](https://github.com/feiskyer))
* `kubectl api-resources` now supports filtering to resources supporting specific verbs, and can output fully qualified resource names suitable for combining with commands like `kubectl get` ([#63254](https://github.com/kubernetes/kubernetes/pull/63254), [@liggitt](https://github.com/liggitt))