* Logging Agent Improvements
    * Bump logging agent version to 0.2-1.5.33-1-k8s-1.
    * Appropriately set log severity for k8s_container.
    * Fix detect exceptions plugin to analyze message field instead of log field.
    * Fix detect exceptions plugin to analyze streams based on local resource id.
    * Disable the metadata agent for monitored resource construction in logging.

*   The default `auto-detect` behavior of the kubelet's `--cloud-provider` flag is removed.
    *   You can manually set `--cloud-provider=auto-detect`, but be aware that this behavior will be removed completely in a future version.

### Other notable changes

* Connections from Pods to Services with 0 endpoints will now ICMP reject immediately, rather than blackhole and timeout. ([#72534](https://github.com/kubernetes/kubernetes/pull/72534), [@thockin](https://github.com/thockin))

- Fixes a rare race condition handling requests that timeout ([#107460](https://github.com/kubernetes/kubernetes/pull/107460), [@liggitt](https://github.com/liggitt)) [SIG API Machinery]
- Mount-utils: Detect potential stale file handle ([#107040](https://github.com/kubernetes/kubernetes/pull/107040), [@andyzhangx](https://github.com/andyzhangx)) [SIG Storage]

### Other (Cleanup or Flake)

- Fixed missing delete events on informer re-lists to ensure all delete events were correctly emitted and using the latest known object state, so that all event handlers and stores always reflect the actual apiserver state as best as possible ([#115620](https://github.com/kubernetes/kubernetes/pull/115620), [@odinuge](https://github.com/odinuge))

HorizontalPodAutoscaler added a `spec.behavior` field that allows scale behavior to be configured. Behaviors are specified separately for scaling up and down. In each direction a stabilization window can be specified as well as a list of policies and how to select amongst them. Policies can limit the absolute number of pods added or removed, or the percentage of pods added or removed. ([#74525](https://github.com/kubernetes/kubernetes/pull/74525), [@gliush](https://github.com/gliush)) [SIG API Machinery, Apps,...

### Other notable changes

* Connections from Pods to Services with 0 endpoints will now ICMP reject immediately, rather than blackhole and timeout. ([#72534](https://github.com/kubernetes/kubernetes/pull/72534), [@thockin](https://github.com/thockin))

* kubelet: Fix bug where `runAsUser: MustRunAsNonRoot` strategy didn't reject a pod with a non-numeric `USER`. ([#56708](https://github.com/kubernetes/kubernetes/pull/56708), [@php-coder](https://github.com/php-coder))
* Add iptables rules to allow Pod traffic even when default iptables policy is to reject. ([#52569](https://github.com/kubernetes/kubernetes/pull/52569), [@tmjd](https://github.com/tmjd))

- The deprecated flag `--lock-object-namespace` and `--lock-object-name` have been removed from kube-scheduler. Please use `--leader-elect-resource-namespace` and `--leader-elect-resource-name` or ComponentConfig instead to configure those parameters. ([#119130](https://github.com/kubernetes/kubernetes/pull/119130), [@SataQiu](https://github.com/SataQiu)) [SIG Scheduling]

## API Changes
- CustomResourceDefinitions now validate documented API semantics of `x-kubernetes-list-type` and `x-kubernetes-map-type` atomic to reject non-atomic sub-types. ([#84722](https://github.com/kubernetes/kubernetes/pull/84722), [@sttts](https://github.com/sttts))