* Kubelet: Add image cache. ([#38375](https://github.com/kubernetes/kubernetes/pull/38375), [@Random-Liu](https://github.com/Random-Liu))
* Collect logs for dead kubelets too ([#37671](https://github.com/kubernetes/kubernetes/pull/37671), [@mtaufen](https://github.com/mtaufen))



# v1.4.7

* Upgrade Stackdriver Logging Agent addon image to 0.6-1.6.0-1 to use Fluentd v1.2. This provides nanoseconds timestamp granularity for logs. ([#70954](https://github.com/kubernetes/kubernetes/pull/70954), [@qingling128](https://github.com/qingling128))
* [GCE] Filter out spammy audit logs from cluster autoscaler. ([#70696](https://github.com/kubernetes/kubernetes/pull/70696), [@loburm](https://github.com/loburm))

* fix azure disk naming matching issue due to case sensitive comparison ([#81720](https://github.com/kubernetes/kubernetes/pull/81720), [@andyzhangx](https://github.com/andyzhangx))
* Fix `kubectl logs -f` for windows server containers. ([#81747](https://github.com/kubernetes/kubernetes/pull/81747), [@Random-Liu](https://github.com/Random-Liu))

## Version 3.12.13

_2021-01-30_

 *  Fix: Work around a crash in Android 10 and 11 that may be triggered when two threads
    concurrently close an SSL socket. This would have appeared in crash logs as
    `NullPointerException: bio == null`.


## Version 3.12.12

_2020-05-17_

 *  Fix: Don't crash when running as a plugin in Android Studio Canary 4.1. To enable

	}

	// Invalid request conditions:
	//
	//   Cannot have both forceStart and forceStop in the same
	//   request; If clientToken is provided, request can only be
	//   to continue receiving logs, so it cannot be start or
	//   stop;
	if (hip.forceStart && hip.forceStop) ||
		(hip.clientToken != "" && (hip.forceStart || hip.forceStop)) {
		err = ErrInvalidRequest
		return
	}

* remove iSCSI volume storage cleartext secrets in logs ([#81215](https://github.com/kubernetes/kubernetes/pull/81215), [@zouyee](https://github.com/zouyee))

		select {
		case logCh <- *b:
		default:
			consoleLogRPC.PutResponse(b)
			// Do not block on slow receivers.
		}
		return nil
	})
}

// ConsoleLog - sends request to peer nodes to get console logs
func (client *peerRESTClient) ConsoleLog(ctx context.Context, kind madmin.LogMask, logCh chan<- []byte) {
	go func() {
		for {
			client.doConsoleLog(ctx, kind, logCh)
			select {
			case <-ctx.Done():

    client.networkInterceptors().add(loggingInterceptor);
    ```

    **Warning:** Avoid `Level.HEADERS` and `Level.BODY` in production because
    they could leak passwords and other authentication credentials to insecure
    logs.

 *  **WebSocket API now uses `RequestBody` and `ResponseBody` for messages.**
    This is a backwards-incompatible API change.

 *  **The DNS service is now pluggable.** In some situations this may be useful

* Add Simplified Chinese translation for kubectl ([#45573](https://github.com/kubernetes/kubernetes/pull/45573), [@shiywang](https://github.com/shiywang))

### **kubectl (CLI)**
* Features

  * `kubectl logs` supports specifying a container name when using label selectors ([#44282](https://github.com/kubernetes/kubernetes/pull/44282), [@derekwaynecarr](https://github.com/derekwaynecarr))

* Upgrade Stackdriver Logging Agent addon image to 0.6-1.6.0-1 to use Fluentd v1.2. This provides nanoseconds timestamp granularity for logs. ([#70954](https://github.com/kubernetes/kubernetes/pull/70954), [@qingling128](https://github.com/qingling128))