- Fixed a bug where the kubelet ephemerally failed with `failed to initialize top level QOS containers: root container [kubepods] doesn't exist`, due to the cpuset cgroup being deleted on cgroup v2 with systemd cgroup manager.
   ([#125923](https://github.com/kubernetes/kubernetes/pull/125923), [@haircommander](https://github.com/haircommander)) [SIG Node and Testing]

```

* Fix fetching pids running in a cgroup, which caused problems with OOM score adjustments & setting the /system cgroup ("misc" in the summary API). ([#36614](https://github.com/kubernetes/kubernetes/pull/36614), [@timstclair](https://github.com/timstclair))

        * Whether all containers should be non-root
        * Supplemental Groups
        * FSGroup - a special supplemental group
        * SELinux options
     * If a pod defines an FSGroup, that Pod’s system (emptyDir, secret, configMap,
etc) volumes and block-device volumes will be owned by the FSGroup, and each
container in the pod will run with the FSGroup as a supplemental group

- Kube-controller-manager: alpha gauge metrics for informer queue length are now published as `informer_queued_items{name=kube-controller-manager,group=<group>,resource=<resource>,version=<version>} <count>` ([#135782](https://g...

### SIG Node

- Fixed a bug in previous releases where a pod could be placed inside another pod's cgroup when specifying --cgroup-root ([#70678](https://github.com/kubernetes/kubernetes/pull/70678), [@dashpole](https://github.com/dashpole))

- When specifying `--(kube|system)-reserved-cgroup`, with `--cgroup-driver=systemd`, it is now possible to use the fully qualified cgroupfs name (i.e. `/test-cgroup.slice`). ([#78793](https://github.com/kubernetes/kubernetes/pull/78793), [@mattjmcnaughton](https://github.com/mattjmcnaughton))

- Promoted CRI communication of the cgroup driver mechanism to beta. The `KubeletCgroupDriverFromCRI` feature gate is now in beta and enabled by default. This allows the kubelet to query the container runtime using CRI to determine the mechanism for cgroup management. If the container runtime doesn't support this, the kubelet falls back to using the configuration file (you can also use the deprecated `--cgroup-driver` command line argument). ([#125828](https://github.com...

  For metrics `etcd_request_duration_seconds`, `etcd_requests_total` and `etcd_request_errors_total` replace `type` label with API `resource` and `group` label.
  For metric `apiserver_selfrequest_total` add a API `group` label.

* Fixes a bug in previous releases where a pod could be placed inside another pod's cgroup when specifying --cgroup-root ([#70678](https://github.com/kubernetes/kubernetes/pull/70678), [@dashpole](https://github.com/dashpole))

  Fixed skip `not found` nodes when reconciling LB backend address pools ([#105188](https://github.com/kubernetes/kubernetes/pull/105188), [@nilo19](https://github.com/nilo19))
- Fixed occasional pod cgroup freeze when using cgroup v1 and systemd driver. ([#104528](https://github.com/kubernetes/kubernetes/pull/104528), [@kolyshkin](https://github.com/kolyshkin))