* to hang (possibly too quickly, but too-quick failures should be very unlikely, given that we
   * used to bail after 20ms during the expected-successful tests, and there we saw a failure rate
   * of ~1/5000, meaning that the other thread's get() call nearly always completes within 20ms if
   * it's going to complete at all).
   *

* Refactor to simplify the hard-traveled path of the KubeletConfiguration object ([#29216](https://github.com/kubernetes/kubernetes/pull/29216), [@mtaufen](https://github.com/mtaufen))
* Fix overflow issue in controller-manager rate limiter ([#31396](https://github.com/kubernetes/kubernetes/pull/31396), [@foxish](https://github.com/foxish))

Please see the [Releases Page](https://github.com/kubernetes/kubernetes/releases) for older releases.

* Fix a scalability issue where high rates of event writes degraded etcd performance. ([#64539](https://github.com/kubernetes/kubernetes/pull/64539), [@ccding](https://github.com/ccding))

A security issue was discovered in Kubernetes that could allow  Windows workloads to run as `ContainerAdministrator` even when those workloads set the `runAsNonRoot` option to `true `.

This issue has been rated low and assigned CVE-2021-25749

### Am I vulnerable?

All Kubernetes clusters with following versions, running Windows workloads with `runAsNonRoot` are impacted

#### Affected Versions

- Implemented logging for failed transactions and the full table in `kube-proxy` with 
`nftables` when using log level 4 or higher. Logging is rate-limited to one entry every 24 hours 
to avoid performance issues. ([#128886](https://github.com/kubernetes/kubernetes/pull/128886), [@npinaeva](https://github.com/npinaeva))
- Implemented the `scheduler_cache_size` metric.

- Fixed a bug that caused a DaemonSet rolling update to hang when its pod gets stuck at terminating.  ([#77773](https://github.com/kubernetes/kubernetes/pull/77773), [@DaiHao](https://github.com/DaiHao))
- Route controller now respects rate limiting to the cloud provider on deletion; previously it was only for create. ([#78581](https://github.com/kubernetes/kubernetes/pull/78581), [@andrewsykim](https://github.com/andrewsykim))

- Fixed clearing of rate-limiter for the queue of checks for cleaning stale pod disruption conditions. The bug could result in the PDB synchronization updates firing too often or the pod disruption cleanups taking too long to happen. ([#114770](https://github.com/kubernet...

- Faster mount detection for linux kernel 5.10+ using openat2 speeding up pod churn rates. On Kernel versions less 5.10, it will fallback to using the original way of detecting mount points i.e by parsing /proc/mounts. ([#109217](https://github.com/kubernetes/kubernetes/pull/109217), [@manugupt1](https://github.com/manugupt1))

- client-go discovery clients constructed using `NewDiscoveryClientForConfig` or `NewDiscoveryClientForConfigOrDie` default to rate limits that cause normal discovery request patterns to take several seconds. This is fixed in https://issue.k8s.io/86168 and will be resolved in v1.17.1. As a workaround, the `Burst` value can be adjusted higher in the rest.Config passed into `New...

- Openapi-controller: remove the trailing `1` character literal from the rate limiting metric `APIServiceOpenAPIAggregationControllerQueue1` and rename it to `open_api_aggregation_controller` to adhere to Prometheus best practices. ([#77979](https://github.com/kubernetes/kubernetes/pull/77979), [@s-urbaniak](https://github.com/s-urbaniak))...