- APIServerTracing now collects spans from etcd client calls, and propagates context to etcd. ([#103216](https://github.com/kubernetes/kubernetes/pull/103216), [@dashpole](https://github.com/dashpole)) [SIG API Machinery, Cloud Provider and Instrumentation]

- The kubelet now only collects cgroups metrics for the node, container runtime, kubelet, pods, and containers. ([#72787](https://github.com/kubernetes/kubernetes/pull/72787), [@dashpole](https://github.com/dashpole))

#### Storage

#### Garbage Collector
* The garbage collector now supports custom APIs added via Custom Resource Definitions
  or aggregated API servers. The garbage collector controller refreshes periodically.
  Therefore, expect a latency of about 30 seconds between when an API is added and when
  the garbage collector starts to manage it.


#### Monitoring/Prometheus

* Kubelet will set extended resource capacity to zero after it restarts. If the extended resource is exported by a device plugin, its capacity will change to a valid value after the device plugin re-connects with the Kubelet. If the extended resource is exported by an external component through direct node status capacity patching, the component should repatch the field after kubelet becomes ready again. During the time gap, pods previously assigned...

* The garbage collector now supports CustomResourceDefinitions and APIServices. ([#65918](https://github.com/kubernetes/kubernetes/pull/65918), [@nikhita](https://github.com/nikhita))

      - [GCE](#gce)
      - [GKE](#gke)
      - [vSphere](#vsphere)
    - [Federation](#federation-2)
      - [kubefed](#kubefed)
      - [Other Notable Changes](#other-notable-changes-13)
    - [Garbage Collector](#garbage-collector)
    - [kubeadm](#kubeadm-2)
    - [kubectl](#kubectl-1)
      - [New Commands](#new-commands)
      - [Create subcommands](#create-subcommands)

- Fixed a bug in the garbage collector controller which could block indefinitely due to a cache sync failure. This fix allows the garbage collector to eventually continue garbage collecting other resources if a given resource cannot be listed or watched. Any objects in the unsynced resource type with owner references with `blockOwnerDeletion: true` will not be known to the garbage collector. Use of `blockOwnerDeletion` has always been best-effort and racy on...

* bazel deb package bugfix: The kubeadm deb package now reloads the kubelet after installation ([#65554](https://github.com/kubernetes/kubernetes/pull/65554), [@rdodev](https://github.com/rdodev))
* The garbage collector now supports CustomResourceDefinitions and APIServices. ([#65915](https://github.com/kubernetes/kubernetes/pull/65915), [@nikhita](https://github.com/nikhita))

- Added `ImageMaximumGCAge` field to Kubelet configuration, which allows a user to set the maximum age an image is unused before it's garbage collected. ([#121275](https://github.com/kubernetes/kubernetes/pull/121275), [@haircommander](https://github.com/haircommander))
- Added `UserNamespacesPodSecurityStandards` feature gate to enable user namespace support for Pod Security Standards.

* fluentd-gcp addon: Fix a bug in the event-exporter, when repeated events were not sent to Stackdriver. ([#50511](https://github.com/kubernetes/kubernetes/pull/50511), [@crassirostris](https://github.com/crassirostris))
* Collect metrics from Heapster in Stackdriver mode. ([#50517](https://github.com/kubernetes/kubernetes/pull/50517), [@piosz](https://github.com/piosz))