- Ensure the `KUBE-MARK-DROP` chain in kube-proxy IPVS mode. The chain is ensured for both IPv4 and IPv6 in dual-stack operation. ([#82214](https://github.com/kubernetes/kubernetes/pull/82214), [@uablrek](https://github.com/uablrek))

* Kubernetes version command line parameter in kubeadm has been updated to drop an unnecessary redirection from ci/latest.txt to ci-cross/latest.txt. Users should know exactly where the builds are stored on Google Cloud storage buckets from now on. For example for 1.9 and 1.10, users can specify ci/latest-1.9 and ci/latest-1.10...

- DRA API: Fixed the `tolerations` field in exact and sub requests to drop properly when the `DRADeviceTaints` API is disabled. ([#132927](https://github.com/kubernetes/kubernetes/pull/132927), [@pohly](https://github.com/pohly))

		meta[xhttp.ContentType] = oi.ContentType
	}

	meta[xhttp.AmzObjectTagging] = oi.UserTags
	meta[xhttp.AmzTagDirective] = "REPLACE"

	if sc == "" {
		sc = oi.StorageClass
	}
	// drop non standard storage classes for tiering from replication
	if sc != "" && (sc == storageclass.RRS || sc == storageclass.STANDARD) {
		meta[xhttp.AmzStorageClass] = sc
	}

	meta[xhttp.MinIOSourceETag] = oi.ETag

with to go1.18, which introduced significant changes to its garbage collection algorithm. As a result, we observed an increase in memory usage for kube-apiserver in larger an heavily loaded clusters up to ~25% (with the benefit of API call latencies drop by up to 10x on 99th percentiles). If the memory increase is not acceptable for you you can mitigate by setting GOGC env variable (for our tests using GOGC=63 brings memory usage back to original value, although the exact value may depend on usage...

- Fixed graceful apiserver shutdown to not drop outgoing bytes before the process terminates. ([#72970](https://github.com/kubernetes/kubernetes/pull/72970), [@sttts](https://github.com/sttts))

### Apps

- Deprecate kubectl top flags related to heapster
  Drop support of heapster in kubectl top ([#87498](https://github.com/kubernetes/kubernetes/pull/87498), [@serathius](https://github.com/serathius)) [SIG CLI]
- Deprecate the `--target-ram-md` flags that is no longer used for anything. ([#91818](https://github.com/kubernetes/kubernetes/pull/91818), [@wojtek-t](https://github.com/wojtek-t)) [SIG API Machinery]

* Fix bug where EndpointSlice controller would attempt to modify shared objects. ([#85368](https://github.com/kubernetes/kubernetes/pull/85368), [@robscott](https://github.com/robscott))
* Revert ensure the KUBE-MARK-DROP chain in kube-proxy mode=iptables. Fix a bug in which kube-proxy deletes the rules associated with the chain in iptables mode. ([#85527](https://github.com/kubernetes/kubernetes/pull/85527), [@aojea](https://github.com/aojea))

* Kubernetes version command line parameter in kubeadm has been updated to drop an unnecessary redirection from ci/latest.txt to ci-cross/latest.txt. Users should know exactly where the builds are stored on Google Cloud storage buckets from now on. For example for 1.9 and 1.10, users can specify ci/latest-1.9 and ci/latest-1.10...

  for background information.

### Scheduling

  - Opaque Integer Resources (OIRs) are deprecated and will be removed in
    version 1.9. Extended Resources (ERs) are a drop-in replacement for OIRs. You can use
    any domain name prefix outside of the `kubernetes.io/` domain instead of the
    `pod.alpha.kubernetes.io/opaque-int-resource-` prefix.

## Notable Features