- Added a new alpha API: ClusterTrustBundle (`certificates.k8s.io/v1alpha1`).
  A ClusterTrustBundle may be used to distribute [X.509](https://www.itu.int/rec/T-REC-X.509) trust anchors to workloads within the cluster. ([#113218](https://github.com/kubernetes/kubernetes/pull/113218), [@ahmedtd](https://github.com/ahmedtd)) [SIG API Machinery, Auth and Testing]

	"github.com/minio/minio/internal/cachevalue"
	xioutil "github.com/minio/minio/internal/ioutil"
	"github.com/minio/minio/internal/logger"
	"github.com/minio/minio/internal/mcontext"
	"github.com/minio/minio/internal/rest"
	"github.com/prometheus/client_golang/prometheus"
	dto "github.com/prometheus/client_model/go"
	"github.com/prometheus/common/expfmt"
	"github.com/prometheus/procfs"
)

- the IP allocator in v1.17.0 can return errors such as `the cluster IP <ip> for service <service-name> is not within the service CIDR <cidr>; please recreate`...

                    && this.sbuf[7] == (byte) 'B') {
                this.smb2 = true;
                // also read the rest of the header
                final int lenDiff = Smb2Constants.SMB2_HEADER_LENGTH - SmbConstants.SMB1_HEADER_LENGTH;
                if (readn(this.in, this.sbuf, 4 + SmbConstants.SMB1_HEADER_LENGTH, lenDiff) < lenDiff) {

    long addressAsLong = 0;
    if (hasEmbeddedIPv4ClientAddress(ip6)) {
      addressAsLong = getEmbeddedIPv4ClientAddress(ip6).hashCode();
    } else {
      // Just extract the high 64 bits (assuming the rest is user-modifiable).
      addressAsLong = ByteBuffer.wrap(ip6.getAddress(), 0, 8).getLong();
    }

    // Many strategies for hashing are possible. This might suffice for now.

* Retry Pod/RC updates in kubectl rolling-update ([#27509](https://github.com/kubernetes/kubernetes/pull/27509), [@janetkuo](https://github.com/janetkuo))
* GCE provider: Create TargetPool with 200 instances, then update with rest ([#27865](https://github.com/kubernetes/kubernetes/pull/27865), [@zmerlynn](https://github.com/zmerlynn))

	// setup becoming a reality we must try to shard the work properly such as
	// pick 10 nodes that precisely can send those 100 requests the first node
	// in the 10 node shard would coordinate between other 9 shards to get the
	// rest of the `99*9` requests.
	//
	// This essentially splits the workload properly and also allows for network
	// utilization to be optimal, instead of blindly throttling the way we are

### Taints and Tolerations
* Add a manager to NodeController that is responsible for removing Pods from Nodes tainted with NoExecute Taints. This feature is beta (as the rest of taints) and enabled by default. It's gated by controller-manager enable-taint-manager flag. ([#40355](https://github.com/kubernetes/kubernetes/pull/40355), [@gmarek](https://github.com/gmarek))

- DRA drivers: the resource slice controller sometimes didn't react properly when kubelet or someone else deleted a recently created ResourceSlice. It incorrectly assumed that the ResourceSlice still exists and didn't recreate it. ([#132683](https://github.com/kubernetes/kubernetes/pull/132683), [@pohly](https://github.com/pohly))...

  //
  // A consequence of these requirements is that the delegate futures cannot be stored in
  // final fields.
  //
  // For simplicity the rest of this description will discuss Futures.catching since it is the
  // simplest instance, though very similar descriptions apply to many other classes in this file.
  //