- Reduced get PV request from KCM pv-controller for CSI volumes ([#134290](https://github.com/kubernetes/kubernetes/pull/134290), [@huww98](https://github.com/huww98)) [SIG Apps and Storage]

        A source is specified by a {@code directory} element, which is relative to the POM.
        The directory content can optionally be reduced to a subset with the {@code includes} and
        {@code excludes} elements. The kind of sources (codes, resources, <i>etc.</i>) and their
        usage (main code, tests, <i>etc.</i>) is specified by the {@code scope} element.

* adjusted http/2 buffer sizes for apiservers to prevent starvation issues between concurrent streams ([#67902](https://github.com/kubernetes/kubernetes/pull/67902), [@liggitt](https://github.com/liggitt))
* Reduced excessive logging from fluentd-gcp-scaler. ([#68837](https://github.com/kubernetes/kubernetes/pull/68837), [@x13n](https://github.com/x13n))

- Fixed an EndpointSlice Controller hashing bug that could cause EndpointSlices to incorrectly handle Pods with duplicate IP addresses. For example this could happen when a new Pod reused an IP that was also assigned to a Pod in a completed state. ([#115907](https://github.com/kubernetes/kubernetes/pull/115907), [@qinqon](https://github.com/qinqon)) [SIG Apps and Network]

- GCE/Windows: When the service cannot be stopped Stackdriver logging processes are now force killed ([#77378](https://github.com/kubernetes/kubernetes/pull/77378), [@yujuhong](https://github.com/yujuhong))
- Reduced GCE log rotation check from 1 hour to every 5 minutes.  Rotation policy is unchanged (new day starts, log file size > 100MB). ([#76352](https://github.com/kubernetes/kubernetes/pull/76352), [@jpbetz](https://github.com/jpbetz))

- Kubeadm: fix the bug that configurable KubernetesVersion not respected during kubeadm join ([#111024](https://github.com/kubernetes/kubernetes/pull/111024), [@SataQiu](https://github.com/SataQiu)) [SIG Cluster Lifecycle]
- Reduced time taken to sync proxy rules on Windows kube-proxy with kernelspace mode ([#110701](https://github.com/kubernetes/kubernetes/pull/110701), [@daschott](https://github.com/daschott)) [SIG Network and Windows]

* Allows changing nodeName in endpoint update. ([#68575](https://github.com/kubernetes/kubernetes/pull/68575), [@prameshj](https://github.com/prameshj))
* Reduced excessive logging from fluentd-gcp-scaler. ([#68837](https://github.com/kubernetes/kubernetes/pull/68837), [@x13n](https://github.com/x13n))

- Scores from PodTopologySpreading have reduced differentiation as maxSkew increases. ([#90820](https://github.com/kubernetes/kubernetes/pull/90820), [@alculquicondor](https://github.com/alculquicondor)) [SIG Scheduling]

- Kubeadm: fix the bug that configurable KubernetesVersion not respected during kubeadm join ([#111022](https://github.com/kubernetes/kubernetes/pull/111022), [@SataQiu](https://github.com/SataQiu)) [SIG Cluster Lifecycle]
- Reduced time taken to sync proxy rules on Windows kube-proxy with kernelspace mode ([#110702](https://github.com/kubernetes/kubernetes/pull/110702), [@daschott](https://github.com/daschott)) [SIG Network and Windows]

- An inefficient lock in EndpointSlice controller metrics cache has been reworked. Network programming latency may be significantly reduced in certain scenarios, especially in clusters with a large number of Services. ([#107091](https://github.com/kubernetes/kubernetes/pull/107091), [@robscott](https://github.com/robscott))