incorrectly reject pods with OutOfCpu errors if they were rapidly scheduled after other pods were reported as complete in the API. The Kubelet now waits to report the phase of a pod as terminal in the API until all running containers are guaranteed to have stopped and no new containers can be started.  Short-lived pods may take slightly longer (~1s) to report Succeeded or Failed after this change. ([#108749](https://github.com/kubernetes/kubernetes/pull/108749), [@bobbypage](https://github.com/bobbypage))...

- Fixed admission metrics histogram bucket sizes to cover 25ms...

Service Level Indicators (SLIs) and Service Level Objectives (SLOs) for the system.
Here's the release [scalability validation report].

[SIG Scalability]: https://github.com/kubernetes/community/tree/master/sig-scalability
[scalability validation report]: https://github.com/kubernetes/sig-release/blob/master/releases/release-1.8/scalability_validation_report.md

### SIG Scheduling

- **Apps**
  - [stable] When replica sets cannot create pods, they will now report detail via the API about the underlying reason ([kubernetes/features#120](https://github.com/kubernetes/enhancements/issues/120))

- Kube-apiserver: Authorization webhooks now report the following metrics:
  - apiserver_authorization_webhook_evaluations_total
  - apiserver_authorization_webhook_duration_seconds

During server startup, prior to the server reporting healthy (via `/healthz`), not all API groups may be reported.
Wait for the server to report healthy (via `/healthz`) before depending on the information provided by the discovery APIs.

		return ErrRequestTimedout
	}

	// Only return ErrClientDisconnected if the provided context is actually canceled.
	// This way downstream context.Canceled will still report ErrRequestTimedout
	if contextCanceled(ctx) && errors.Is(ctx.Err(), context.Canceled) {
		return ErrClientDisconnected
	}

	// Unwrap the error first
	err = unwrapAll(err)

	switch err {

		}
		// Hadoop makes the max-keys=2 listing call just to find if the directory is empty or not, or in the case
		// of an object to check for object existence. For versioned buckets, MinIO's non-recursive
		// call will report top level prefixes in deleted state, whereas spark/hadoop interpret this as non-empty
		// and throw a 404 exception. This is especially a problem for spark jobs overwriting the same partition

- Terminating a restartPolicy=Never pod no longer has a chance to report the pod succeeded when it actually failed. ([#88440](https://github.com/kubernetes/kubernetes/pull/88440), [@smarterclayton](https://github.com/smarterclayton)) [SIG Node and Testing]

- Generating informers for packages in directories containing dots in their name is broken. This will be fixed in v1.16.1. ([#82860](https://github.com/kubernetes/kubernetes/issues/82860))
- kube-scheduler won't be able to report scheduling Events if `events.k8s.io/v1beta1` API is disabled. We are targeting the fix for v1.16.2 ([#83203](https://github.com/kubernetes/kubernetes/issues/83203))