- Fake clientsets now use a common, generic implementation. The corresponding structs are now private; callers must use the corresponding constructors. ([#126503](https://github.com/kubernetes/kubernetes/pull/126503), [@skitt](https://github.com/skitt)) [SIG API Machinery, Architecture, Auth and Instrumentation]

### Alpha

- Introduce a new admission controller for RuntimeClass. Initially, RuntimeClass will be used to apply the pod overhead associated with a given RuntimeClass to the Pod `spec` if a corresponding RuntimeClassName is specified. PodOverhead is an alpha feature as of Kubernetes 1.16. ([#78484](https://github.com/kubernetes/kubernetes/pull/78484), [@egernst](https://github.com/egernst))

### CVE-2025-5187: Nodes can delete themselves by adding an OwnerReference

### CVE-2025-5187: Nodes can delete themselves by adding an OwnerReference

- The following features are unconditionally enabled and the corresponding `--feature-gates` flags have been removed: `PodPriority`, `TaintNodesByCondition`, `ResourceQuotaScopeSelectors` and `ScheduleDaemonSetPods` ([#86210](https://github.com/kubernetes/kubernetes/pull/86210), [@draveness](https://github.com/draveness))...

* Fix a bug in the IPVS proxier where virtual servers are not cleaned up even though the corresponding Service object was deleted. ([#80942](https://github.com/kubernetes/kubernetes/pull/80942), [@gongguan](https://github.com/gongguan))

- Fix a bug on the endpoint slices mirroring controller where endpoint NotReadyAddresses were mirrored as Ready to the corresponding EndpointSlice ([#102683](https://github.com/kubernetes/kubernetes/pull/102683), [@aojea](https://github.com/aojea)) [SIG Apps and Network]

- Take the context as the first argument of Schedule. ([#82119](https://github.com/kubernetes/kubernetes/pull/82119), [@wgliang](https://github.com/wgliang))
- Fixed an issue that the correct PluginConfig.Args is not passed to the corresponding PluginFactory in kube-scheduler when multiple PluginConfig items are defined. ([#82483](https://github.com/kubernetes/kubernetes/pull/82483), [@everpeace](https://github.com/everpeace))

- Setting a mirror pod's phase to Succeeded or Failed can prevent the corresponding static pod from restarting due mutation of a Kubelet cache. ([#116482](https://github.com/kubernetes/kubernetes/pull/116482), [@smarterclayton](https://github.com/smarterclayton)) [SIG Node]

### Other (Cleanup or Flake)

*   You can now override the health check parameters for AWS ELBs by specifying annotations on the corresponding service. The new annotations are: `healthy-threshold`, `unhealthy-threshold`, `timeout`, `interval`. The prefix for all annotations is  `service.beta.kubernetes.io/aws-load-balancer-healthcheck-`. ([#56024](https://github.com/kubernetes/kubernetes/pull/56024),[...