- Migrated the ResourceQuota controller (within `kube-controller-manager`) to use [contextual logging](https://k8s.io/docs/concepts/cluster-administration/system-logs/#contextual-logging). ([#113315](https://github.com/kubernetes/kubernetes/pull/113315), [@ncdc](https://github.com/ncdc)) [SIG API Machinery, Apps and Testing]

- Kubelet: Fixes memory usage within the kubelet's inotify subsystem. ([#108335](https://github.com/kubernetes/kubernetes/pull/108335), [@rphillips](https://github.com/rphillips)) [SIG API Machinery, CLI, Cloud Provider, Cluster Lifecycle, Instrumentation, Node and Storage]

* Fix in vSphere Cloud Provider to report disk is detach when VM is not found. ([#62220](https://github.com/kubernetes/kubernetes/pull/62220), [@abrarshivani](https://github.com/abrarshivani))
* Restores ability to enable/disable resource within an API GroupVersion individually. ([#62116](https://github.com/kubernetes/kubernetes/pull/62116), [@deads2k](https://github.com/deads2k))

Added an `--endpoint-reconciler-type` command-line argument to select the endpoint reconciler to use. The default is to use the 'master-count' reconciler which is the default for 1.9 and in use prior to 1.9. The 'lease' reconciler stores endpoints within the storage api for better cleanup of deleted (or removed) API servers. The 'none' reconciler is a no-op reconciler, which can be used in self-hosted environments.   ([#51698](https://github.com/kubernetes/kubernetes/pull/51698), [@rphillips](htt...

- Migrated the Job controller (within `kube-controller-manager`) to use [contextual logging](https://k8s.io/docs/concepts/cluster-administration/system-logs/#contextual-logging). ([#116910](https://github.com/kubernetes/kubernetes/pull/116910), [@fatsheep9146](https://github.com/fatsheep9146)) [SIG API Machinery, Apps and Testing]

#### Other deprecations:
- The k8s.io/node-api component is no longer updated. Instead, use the RuntimeClass types located within k8s.io/api, and the generated clients located within k8s.io/client-go ([#87503](https://github.com/kubernetes/kubernetes/pull/87503), [@liggitt](https://github.com/liggitt)) [SIG Node and Release]

multiple clusters ensures that cluster-level autoscalers are
consistently deployed across those clusters, and dynamically managed
to ensure that autoscaling can occur optimially in all clusters,
within a set of global constraints on the total number of replicas
permitted across all clusters.  If replicas are not
required in some clusters due to low system load or insufficient quota

- The PodShareProcessNamespace feature to configure PID namespace sharing within a pod has been promoted to beta. ([#66507](https://github.com/kubernetes/kubernetes/pull/66507), [@verb](https://github.com/verb))

## API Changes

- the IP allocator in v1.17.0 can return errors such as `the cluster IP <ip> for service <service-name> is not within the service CIDR <cidr>; please recreate` in the logs of the kube-apiserver. The cause is incorrect CIDR calculations if the service CIDR (`--service-cluster-ip-range`) is set to bits lower than `/16`. This is fixed in http://issue.k8s.io/86534 and...

* APIServices with kube-like versions (e.g. `v1`, `v2beta1`, etc.) will be sorted appropriately within each group.  ([#64004](https://github.com/kubernetes/kubernetes/pull/64004), [@mbohlool](https://github.com/mbohlool))