* `federation/deploy/deploy.sh` was an interim solution introduced in Kubernetes v1.4 to simplify the federation control plane deployment experience. Now that we have `kubefed`, we are deprecating `deploy.sh` scripts. ([#38902](https://github.com/kubernetes/kubernetes/pull/38902), [@madhusudancs](https://github.com/madhusudancs))

* `metrics-server` is now the recommended way to provide the resource
  metrics API. Deploy `metrics-server` as an add-on in the same way that you deploy Heapster.

##### Cluster Autoscaler

* Cluster autoscaler is now GA
* Cluster support size is increased to 1000 nodes
* Respect graceful pod termination of up to 10 minutes

- cloud.google.com/go/dataplex: v1.6.0 → v1.9.0
- cloud.google.com/go/dataqna: v0.7.0 → v0.8.1
- cloud.google.com/go/datastore: v1.11.0 → v1.13.0
- cloud.google.com/go/datastream: v1.7.0 → v1.10.0
- cloud.google.com/go/deploy: v1.8.0 → v1.13.0
- cloud.google.com/go/dialogflow: v1.32.0 → v1.40.0
- cloud.google.com/go/dlp: v1.9.0 → v1.10.1
- cloud.google.com/go/documentai: v1.18.0 → v1.22.0
- cloud.google.com/go/domains: v0.8.0 → v0.9.1

- The in-tree GCE PD plugin `kubernetes.io/gce-pd` is now deprecated and will be removed in 1.21. Users that self-deploy Kubernetes on GCP should enable CSIMigration + CSIMigrationGCE features and install the GCE PD CSI Driver (https://github.com/kubernetes-sigs/gcp-compute-persistent-disk-csi-driver) to avoid disruption to existing Pod and PVC objects at that time....

    - [Container Storage Interface graduations](#container-storage-interface-graduations)
    - [SIG Windows development tools](#sig-windows-development-tools)
    - [Deploy a more secure control plane with kubeadm](#deploy-a-more-secure-control-plane-with-kubeadm)
    - [etcd moves to version 3.5.0](#etcd-moves-to-version-350)
    - [Kubernetes Node system swap support](#kubernetes-node-system-swap-support)

### Cluster Lifecycle

- cloud.google.com/go/dataproc/v2: v2.3.0
- cloud.google.com/go/dataqna: v0.8.4
- cloud.google.com/go/datastore: v1.15.0
- cloud.google.com/go/datastream: v1.10.3
- cloud.google.com/go/deploy: v1.17.0
- cloud.google.com/go/dialogflow: v1.48.1
- cloud.google.com/go/dlp: v1.11.1
- cloud.google.com/go/documentai: v1.23.7
- cloud.google.com/go/domains: v0.9.4
- cloud.google.com/go/edgecontainer: v1.1.4

- cloud.google.com/go/dataproc/v2: v2.0.1 → v2.3.0
- cloud.google.com/go/dataqna: v0.8.1 → v0.8.4
- cloud.google.com/go/datastore: v1.13.0 → v1.15.0
- cloud.google.com/go/datastream: v1.10.0 → v1.10.3
- cloud.google.com/go/deploy: v1.13.0 → v1.17.0
- cloud.google.com/go/dialogflow: v1.40.0 → v1.48.1
- cloud.google.com/go/dlp: v1.10.1 → v1.11.1
- cloud.google.com/go/documentai: v1.22.0 → v1.23.7
- cloud.google.com/go/domains: v0.9.1 → v0.9.4

* kubeadm now accepts `--apiserver-extra-args`, `--controller-manager-extra-args` and `--scheduler-extra-args`, making it possible to override / specify additional flags for control plane components. One good example is to deploy Kubernetes with a different admission-control flag on API server. ([#58080](https://github.com/kubernetes/kubernetes/pull/58080), [@simonferquel](https://github.com/simonferquel))

- CoreDNS is only officially supported on Linux at this time.  As such, when kubeadm is used to deploy this component into your kubernetes cluster, it will be restricted (using `nodeSelectors`) to run only on nodes with that operating system. This ensures that in clusters which include Windows nodes, the scheduler will not ever attempt to place CoreDNS...