* fixes a memory leak in the kube-controller-manager observed when large numbers of pods with tolerations are created/deleted ([#65339](https://github.com/kubernetes/kubernetes/pull/65339), [@liggitt](https://github.com/liggitt))

- Kubeadm: modify the etcd static Pod liveness and readyness probes to use a new etcd 3.5.3+ HTTP(s) health check endpoint "/health?serializable=true" that allows to track the health of individual etcd members and not fail all members if a single member is not healthy in the etcd cluster. ([#110072](https://github.com/kubernetes/kubernetes/pull/110072), [@neolit123](https://github.com/neolit123)) [SIG Cluster Lifecycle]

* Upgrade to etcd client 3.3.17 to fix bug where etcd client does not parse IPv6 addresses correctly when members are joining, and to fix bug where failover on multi-member etcd cluster fails certificate check on DNS mismatch ([#83968](https://github.com/kubernetes/kubernetes/pull/83968), [@jpbetz](https://github.com/jpbetz))

- Kubeadm: during "reset" do not remove the only remaining stacked etcd member from the cluster and just proceed with the cleanup of the local etcd storage. ([#91145](https://github.com/kubernetes/kubernetes/pull/91145), [@tnqn](https://github.com/tnqn)) [SIG Cluster Lifecycle]

- If `kubeadm reset` finds no etcd member ID for the peer it removes during the `remove-etcd-member` phase, it continues immediately to other phases, instead of retrying the phase for up to 3 minutes before continuing. ([#118192](https://github.com/kubernetes/kubernetes/pull/118192), [@dlipovetsky](https://github.com/dlipovetsky))...

- Kubeadm: add the flag "--experimental-initial-corrupt-check" to etcd static Pod manifests to ensure etcd member data consistency ([#109076](https://github.com/kubernetes/kubernetes/pull/109076), [@neolit123](https://github.com/neolit123)) [SIG Cluster Lifecycle]

## Dependencies

### Added
_Nothing has changed._

- Kubeadm: add the flag "--experimental-initial-corrupt-check" to etcd static Pod manifests to ensure etcd member data consistency ([#109075](https://github.com/kubernetes/kubernetes/pull/109075), [@neolit123](https://github.com/neolit123)) [SIG Cluster Lifecycle]

## Dependencies

### Added
_Nothing has changed._

* If the cluster node count exceeds the GCE TargetPool maximum (currently 1000), ([#25178](https://github.com/kubernetes/kubernetes/pull/25178), [@zmerlynn](https://github.com/zmerlynn))
    * randomly select which nodes are members of Kubernetes External Load Balancers.
* Clarify supported version skew between masters, nodes, and clients ([#25087](https://github.com/kubernetes/kubernetes/pull/25087), [@ihmccreery](https://github.com/ihmccreery))

## Changelog since v1.15.0

### Other notable changes

* kubeadm: implement support for concurrent add/remove of stacked etcd members ([#79677](https://github.com/kubernetes/kubernetes/pull/79677), [@neolit123](https://github.com/neolit123))

- Moved fluentd-elasticsearch addon images to community controlled location ([#73819](https://github.com/kubernetes/kubernetes/pull/73819), [@coffeepac](https://github.com/coffeepac))
- Removed local etcd members from the etcd cluster when `kubeadm reset` ([#74112](https://github.com/kubernetes/kubernetes/pull/74112), [@pytimer](https://github.com/pytimer))