- [Cluster Autoscaler](#cluster-autoscaler)
    - [Deployment](#deployment)
    - [Federation](#federation)
    - [Internal Storage Layer](#internal-storage-layer)
    - [Node Components](#node-components)
    - [kubectl](#kubectl)
    - [RBAC](#rbac)
    - [Scheduling](#scheduling)
    - [Service](#service)
    - [StatefulSet](#statefulset)
    - [Volumes](#volumes)

- If `PreBindPreFlight` returned `Skip`, the scheduler didn't run the plugin at `PreBind`.
  If any `PreBindPreFlight` returned `Success`, the scheduler put NominatedNodeName to the pod 
  so that other components (such as the cluster autoscaler) could notice the pod was going to be bound to the node. ([#133021](https://github.com/kubernetes/kubernetes/pull/133021), [@sanposhiho](https://github.com/sanposhiho)) [SIG Scheduling and Testing]

    - [**kubectl (CLI)**](#kubectl-cli-1)
    - [**Networking**](#networking)
      - [Network Policy](#network-policy)
      - [Load Balancing](#load-balancing)
    - [**Node Components**](#node-components)
      - [Container Runtime Interface](#container-runtime-interface)
    - [**Scheduling**](#scheduling-1)
      - [Scheduler Extender](#scheduler-extender)
    - [**Storage**](#storage-1)

      - [[alpha] Federated Jobs](#alpha-federated-jobs)
      - [[alpha] Federated Horizontal Pod Autoscaling (HPA)](#alpha-federated-horizontal-pod-autoscaling-hpa)
    - [Node Components](#node-components)
      - [Autoscaling and Metrics](#autoscaling-and-metrics)
        - [Cluster Autoscaler](#cluster-autoscaler)
      - [Container Runtime Interface (CRI)](#container-runtime-interface-cri)

- kubeadm: all master components are now exclusively relying on the `PriorityClassName` pod spec for annotating them as cluster critical components. Since `scheduler.alpha.kubernetes.io/critical-pod` annotation is no longer supported by Kubernetes 1.14 this annotation is no longer added to master components. ([#73857](https://github.com/kubernetes/kubernetes/pull/73857), [@ereslibre](https://github.com/ereslibre))...

  It is possible that this will cause problems with some third-party components
  that improperly depended on those rules. If this affects you, you can run
  `kubelet` with `--feature-gates=IPTablesOwnershipCleanup=false`, but a bug should also be filed against the third-party component. ([#114472](https://github.com/kubernetes/kubernetes/pull/114472), [@danwinship](https://github.com/danwinship))

- Added to `MutableFeatureGate` the ability to override the default setting of feature gates, to allow default-enabling a feature on a component-by-component basis instead of for all affected components simultaneously. ([#122647](https://github.com/kubernetes/kubernetes/pull/122647), [@benluddy](https://github.com/benluddy))

This API provided status of etcd, kube-scheduler, and kube-controller-manager components, but only worked when those components were local to the API server, and when kube-scheduler and kube-controller-manager exposed unsecured health endpoints. Instead of this API, etcd health is included in the kube-apiserver health check and kube-scheduler/kube-controller-manager health checks can be made directly against those components' health endpoints. ([#93570](https://github.com/kubernetes/kubernetes/pull/93570),...

- kubeadm: graduated the WaitForAllControlPlaneComponents feature gate to Beta. When checking the health status of a control plane component, make sure that the address and port defined as arguments in the respective component's static Pod manifest are used. ([#129620](https://github.com/kubernetes/kubernetes/pull/129620), [@neolit123](https://github.com/neolit123)) [SIG Cluster Lifecycle]

* calico-node addon tolerates all NoExecute and NoSchedule taints by default. ([#57122](https://github.com/kubernetes/kubernetes/pull/57122), [@caseydavenport](https://github.com/caseydavenport))
* Allow kubernetes components to react to SIGTERM signal and shutdown gracefully. ([#57756](https://github.com/kubernetes/kubernetes/pull/57756), [@mborsz](https://github.com/mborsz))