* Pass cluster_location argument to Heapster ([#65176](https://github.com/kubernetes/kubernetes/pull/65176), [@kawych](https://github.com/kawych))
* fixes a memory leak in the kube-controller-manager observed when large numbers of pods with tolerations are created/deleted ([#65339](https://github.com/kubernetes/kubernetes/pull/65339), [@liggitt](https://github.com/liggitt))

- Kube-proxy, when using a Service with External or LoadBalancer IPs on UDP services , was consuming a large amount of CPU because it was not filtering by the Service destination port and trying to delete all the UDP entries associated to the service. ([#130505](https://github.com/kubernetes/kubernetes/pull/130505), [@aojea](https://github.com/aojea)) [SIG...

Because the release includes breaking API changes, we're changing the project's
package name from `com.squareup.okhttp` to `okhttp3`. This should make it
possible for large applications to migrate incrementally. The Maven group ID
is now `com.squareup.okhttp3`. For an explanation of this strategy, see Jake
Wharton's post, [Java Interoperability Policy for Major Version
Updates][major_versions].

- Kube-apiserver: `gzip` compression switched from level 4 to level 1 to improve large list call latencies in exchange for higher network bandwidth usage (10-50% higher). This increases the headroom before very large unpaged list calls exceed request timeout limits. ([#112299](https://github.com/kubernetes/kubernetes/pull/112299), [@shyamjvs](https://github.com/shyamjvs))

2B1B..2B1F    ; valid                  ;      ; NV8    # 5.1  BLACK LARGE SQUARE..BLACK PENTAGON
2B20..2B23    ; valid                  ;      ; NV8    # 5.0  WHITE PENTAGON..HORIZONTAL BLACK HEXAGON
2B24..2B4C    ; valid                  ;      ; NV8    # 5.1  BLACK LARGE CIRCLE..RIGHTWARDS ARROW ABOVE REVERSE TILDE OPERATOR

- A new feature was enabled to improve the performance of the iptables mode of `kube-proxy` in large clusters. No action was required, however:
  

  - Speed up binpacking by reducing the number of PreFilter calls (call once per pod instead of #pods*#nodes times)
  - Speed up finding unneeded nodes by 5x+ in very large clusters by reducing the number of PreFilter calls
  - Expose `--max-nodes-total` as a metric
  - Errors in `IncreaseSize` changed from type `apiError` to `cloudProviderError`

* kube-apiserver: a request body of a CREATE/UPDATE/PATCH/DELETE resource operation larger than 100 MB will return a 413 "request entity too large" error. ([#73805](https://github.com/kubernetes/kubernetes/pull/73805), [@caesarxuchao](https://github.com/caesarxuchao))

## Changes by Kind

### Bug or Regression

- Kube-apiserver: gzip compression switched from level 4 to level 1 to improve large list call latencies in exchange for higher network bandwidth usage (10-50% higher). This increases the headroom before very large unpaged list calls exceed request timeout limits. ([#112400](https://github.com/kubernetes/kubernetes/pull/112400), [@shyamjvs](https://github.com/shyamjvs)) [SIG API Machinery]

### Bug or Regression

- 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. ([#107169](https://github.com/kubernetes/kubernetes/pull/107169), [@robscott](https://github.com/robscott)) [SIG Apps and Network]