- Reduced the number of cloud API calls and service downtime caused by excessive re-configurations of cluster LBs with externalTrafficPolicy=Local when node readiness changes (https://github.com/kubernetes/kubernetes/issues/111539). The service controller (in cloud-controller-manager) will avoid resyncing nodes which are transitioning...

    callB.enqueue(callback)
    assertThat(server.takeRequest().url.encodedPath).isEqualTo("/a")
    callback.await(requestA.url).assertBody("A")
    // At this point we know the callback is ready, and that it will receive a cancel failure.
    callback
      .await(requestB.url)
      .assertFailure("canceled", "Canceled", "Socket closed", "Socket is closed")
  }

  @Test
  fun canceledBeforeIOSignalsOnFailure_HTTPS() {

* let patch use --local flag like `kubectl set image` ([#26722](https://github.com/kubernetes/kubernetes/pull/26722), [@deads2k](https://github.com/deads2k))
* enable recursive processing in kubectl edit ([#25085](https://github.com/kubernetes/kubernetes/pull/25085), [@metral](https://github.com/metral))

* A pod’s Quality of Service Class is now available in its Status.
* Upgrade cAdvisor library to v0.25.0. Notable changes include,
  - Container filesystem usage tracking disabled for device mapper due to excessive
    IOPS.
  - Ignore `.mount` cgroups, fixing disappearing stats.
* A new field `terminationMessagePolicy` has been added to containers that allows

- KEP-3857: Recursive Read-only (RRO) mounts: promote to GA ([#130116](https://github.com/kubernetes/kubernetes/pull/130116), [@AkihiroSuda](https://github.com/AkihiroSuda)) [SIG Apps, Node and Testing]

- Reduced excessive logging in the Volume Binding scheduler plugin by lowering the verbosity of high-frequency messages from `V(4)` to `V(5)`. ([#132840](https://github.com/kubernetes/kubernetes/pull/132840), [@ppmechlinski](https://github.com/ppmechlinski)) [SIG Autoscaling,...

			}
			// Since `apiRouter` satisfies `http.Handler` it has a
			// ServeHTTP to execute the logic of the handler.
			apiRouter.ServeHTTP(rec, req)

			if rec.Code != 200 {
				t.Errorf("Test %d: Did not receive a 200 response: %d", i+1, rec.Code)
			}
			contentLength = rec.Header().Get("Content-Length")
		}

		if contentLength != fmt.Sprintf("%d", objectLength(input)) {

* Fixes a flaw (CVE-2019-11253) in json/yaml decoding where large or malformed documents could consume excessive server resources. Request bodies for normal API requests (create/delete/update/patch operations of regular resources) are now limited to 3MB. ([#83261](https://github.com/kubernetes/kubernetes/pull/83261), [@liggitt](https://github.com/liggitt))...

- Scheduler's assumed pods have 2min instead of 30s to receive nodeName pod updates ([#106633](https://github.com/kubernetes/kubernetes/pull/106633), [@ahg-g](https://github.com/ahg-g)) [SIG Scheduling]

## Dependencies

### Added
_Nothing has changed._

### Changed

- Sidecar termination is now serialized and each sidecar container will receive a `SIGTERM` after all main containers and later starting sidecar containers have terminated. ([#120620](https://github.com/kubernetes/kubernetes/pull/120620), [@tzneal](https://github.com/tzneal))