events_with_failures_and_retries@2x.png...

- A new kubelet metric `image_pull_duration_seconds` was added. The metric tracks the duration (in seconds) it takes for an image to be pulled, including the time spent in the waiting queue of image puller. The metric is broken down by bucketed image size. ([#121719](https://github.com/kubernetes/kubernetes/pull/121719), [@ruiwen-zhao](https://github.com/ruiwen-zhao))

	if metric.Description.Type == histogramMetric {
		if metric.Histogram == nil {
			return
		}
		for k, v := range metric.Histogram {
			pmetric, err := prometheus.NewConstMetric(
				prometheus.NewDesc(
					prometheus.BuildFQName(string(metric.Description.Namespace),
						string(metric.Description.Subsystem),

description: Optional. nullable: true type: boolean match: description: Match allows provides the scope of the override. oneOf: - not: anyOf: - required: - metric - required: - customMetric - required: - metric - required: - customMetric properties: customMetric: description: Allows free-form specification of a metric. minLength: 1 type: string metric: description: One of the well-known Istio Standard Metrics. enum: - ALL_METRICS - REQUEST_COUNT - REQUEST_DURATION - REQUEST_SIZE - RESPONSE_SIZE - TCP_OPENED_CONNECTIONS...

  /** The maximum number of segments to allow; used to bound constructor arguments. */
  static final int MAX_SEGMENTS = 1 << 16; // slightly conservative

  /** Number of (unsynchronized) retries in the containsValue method. */
  static final int CONTAINS_VALUE_RETRIES = 3;

  /**
   * Number of cache access operations that can be buffered per segment before the cache's recency

  /** The maximum number of segments to allow; used to bound constructor arguments. */
  static final int MAX_SEGMENTS = 1 << 16; // slightly conservative

  /** Number of (unsynchronized) retries in the containsValue method. */
  static final int CONTAINS_VALUE_RETRIES = 3;

  /**
   * Number of cache access operations that can be buffered per segment before the cache's recency

go.opentelemetry.io/otel/metric v1.26.0 h1:7S39CLuY5Jgg9CrnA9HHiEjGMF/X2VHvoXGgSllRz30=
go.opentelemetry.io/otel/metric v1.26.0/go.mod h1:SY+rHOI4cEawI9a7N1A4nIg/nTQXe1ccCNWYOJUrpX4=
go.opentelemetry.io/otel/sdk v1.26.0 h1:Y7bumHf5tAiDlRYFmGqetNcLaVUZmh4iYfmGxtmz7F8=
go.opentelemetry.io/otel/sdk v1.26.0/go.mod h1:0p8MXpqLeJ0pzcszQQN4F0S5FVjBLgypeGSngLsmirs=

// permanent deletes (by specifying a version ID in the delete operation) have three states "Pending", "Complete"
// and "Failed" to mark the status of the replication of "DELETE" operation. All failed operations can
// then be retried by healing. In the case of permanent deletes, until the replication is completed on the
// target cluster, the object version is marked deleted on the source and hidden from listing. It is permanently

    callback.await(request.url).assertHandshake()
  }

  @Test
  fun recoverWhenRetryOnConnectionFailureIsTrue() {
    // Set to 2 because the seeding request will count down before the retried request does.
    val requestFinished = CountDownLatch(2)
    val dispatcher: QueueDispatcher =
      object : QueueDispatcher() {
        override fun dispatch(request: RecordedRequest): MockResponse {

* Only set sysctls for infra containers ([#32383](https://github.com/kubernetes/kubernetes/pull/32383), [@sttts](https://github.com/sttts))
* fix kubectl taint e2e flake: add retries for removing taint ([#33872](https://github.com/kubernetes/kubernetes/pull/33872), [@kevin-wangzefeng](https://github.com/kevin-wangzefeng))