}

  /**
   * Enable the accumulation of {@link CacheStats} during the operation of the cache. Without this
   * {@link Cache#stats} will return zero for all statistics. Note that recording stats requires
   * bookkeeping to be performed with each operation, and thus imposes a performance penalty on
   * cache operation.
   *
   * @return this {@code CacheBuilder} instance (for chaining)

   * exceeded.
   *
   * The page replacement algorithm's data structures are kept casually consistent with the map. The
   * ordering of writes to a segment is sequentially consistent. An update to the map and recording
   * of reads may not be immediately reflected on the algorithm's data structures. These structures
   * are guarded by a lock and operations are applied in batches to avoid lock contention. The

	* Allow setting KubernetesClusterID or KubernetesClusterTag in combination with VPC. ([#42512](https://github.com/kubernetes/kubernetes/pull/42512), [@scheeles](https://github.com/scheeles))

	* Start recording cloud provider metrics for AWS ([#43477](https://github.com/kubernetes/kubernetes/pull/43477), [@gnufied](https://github.com/gnufied))

- Fixed recording the `kubelet_container_resize_requests_total` metric to include all resize-related updates. ([#133060](https://github.com/kubernetes/kubernetes/pull/133060), [@natasha41575](https://github.com/natasha41575))

    // number 0. Since the client may have attempted to reuse the broken connection just before
    // creating a fresh connection, the server may have recorded 2 requests at this point. The order
    // of recording is non-deterministic.
    val requestAfter = server.takeRequest()
    assertThat(
      requestAfter.exchangeIndex == 0 ||
        server.requestCount == 3 &&
        server.takeRequest().exchangeIndex == 0,

  2. if there's a transient error happened during "provisioning/deletion", i.e., a volume is still in-use while a deleteVolume has been called, original `storage_operation_duration_seconds` will NOT wait until a volume has been finally deleted before reporting an inaccurate latency metric. The newly implemented metric...

- Implemented logging and event recording for probe results with an `Unknown` status in the kubelet's prober module. This helped improve the diagnosis and monitoring of cases where container probes returned an `Unknown` result, enhancing the observability and reliability of health checks....

- DRA: when enabling the scheduler queuing hint feature, pods got stuck as unschedulable for a while unnecessarily because recording the name of the generated ResourceClaim did not trigger scheduling. ([#127497](https://github.com/kubernetes/kubernetes/pull/127497), [@pohly](https://github.com/pohly)) [SIG Auth, Node, Scheduling and Testing]

- Kubelet startup now fails CRI connection if service or image endpoint is throwing any error ([#115102](https://github.com/kubernetes/kubernetes/pull/115102), [@saschagrunert](https://github.com/saschagrunert))
- Kubelet: fix recording issue when pulling image did finish ([#114904](https://github.com/kubernetes/kubernetes/pull/114904), [@TommyStarK](https://github.com/TommyStarK)) [SIG Node]

Paul Merlin <******@****.***> 1688825197 +0200