- To reduce chances of throttling, VM cache is set to nil when Azure node provisioning state is deleting ([#87635](https://github.com/kubernetes/kubernetes/pull/87635), [@feiskyer](https://github.com/feiskyer)) [SIG Cloud Provider]

* azure: Add allow unsafe read from cache ([#83685](https://github.com/kubernetes/kubernetes/pull/83685), [@aramase](https://github.com/aramase))

    executeSynchronously("/", "Cache-Control", "only-if-cached")
      .assertCode(504)
      .assertBody("")
      .assertNoNetworkResponse()
      .assertNoCacheResponse()
  }

  @Test
  fun networkDropsOnConditionalGet() {
    client =
      client.newBuilder()
        .cache(cache)
        .build()

    // Seed the cache.
    server.enqueue(
      MockResponse(

- Ensure Job status updates are batched by 1s. This fixes an unlikely scenario when a sequence of immediately 
  completing pods could trigger a sequence of non-batched Job status updates. ([#118470](https://github.com/kubernetes/kubernetes/pull/118470), [@mimowo](https://github.com/mimowo)) [SIG Apps]

const (
	replicateMetadata replicationAction = "metadata"
	replicateNone     replicationAction = "none"
	replicateAll      replicationAction = "all"
)

// matches k1 with all keys, returns 'true' if one of them matches
func equals(k1 string, keys ...string) bool {
	for _, k2 := range keys {
		if strings.EqualFold(k1, k2) {
			return true
		}
	}
	return false
}

### Network

- Fix the bug PIP's DNS is deleted if no DNS label service annotation isn't set. ([#87312](https://github.com/kubernetes/kubernetes/pull/87312), [@nilo19](https://github.com/nilo19)) [SIG Cloud Provider]
- Fix: set nil cache entry based on old cache ([#87591](https://github.com/kubernetes/kubernetes/pull/87591), [@aramase](https://github.com/aramase)) [SIG Cloud Provider]

    @CheckForNull transient Cache<K, V> delegate;

    ManualSerializationProxy(LocalCache<K, V> cache) {
      this(
          cache.keyStrength,
          cache.valueStrength,
          cache.keyEquivalence,
          cache.valueEquivalence,
          cache.expireAfterWriteNanos,
          cache.expireAfterAccessNanos,
          cache.maxWeight,
          cache.weigher,

  // +optional
  optional string message = 4;
}

// An empty preferred scheduling term matches all objects with implicit weight 0
// (i.e. it's a no-op). A null preferred scheduling term matches no objects (i.e. is also a no-op).
message PreferredSchedulingTerm {
  // Weight associated with matching the corresponding nodeSelectorTerm, in the range 1-100.

  - The disk-cached discovery client is moved from k8s.io/client-go/discovery to k8s.io/client-go/discovery/cached/disk.
The memory-cached discovery client is moved from k8s.io/client-go/discovery/cached to k8s.io/client-go/discovery/cached/memory.
([#72214](https://github.com/kubernetes/kubernetes/pull/72214), [@caesarxuchao](https://github.com/caesarxuchao))