- Fix the `cloudprovider_azure_api_request_duration_seconds` metric buckets to correctly capture the latency metrics. Previously, the majority of the calls would fall in the "+Inf" bucket. ([#95375](https://github.com/kubernetes/kubernetes/pull/95375), [@marwanad](https://github.com/marwanad)) [SIG Cloud Provider and Instrumentation]

import org.opensearch.search.SearchHit;
import org.opensearch.search.SearchHits;
import org.opensearch.search.aggregations.AggregationBuilders;
import org.opensearch.search.aggregations.bucket.filter.FilterAggregationBuilder;
import org.opensearch.search.aggregations.bucket.terms.TermsAggregationBuilder;
import org.opensearch.search.collapse.CollapseBuilder;
import org.opensearch.search.fetch.subphase.highlight.HighlightBuilder;

- Fix the `cloudprovider_azure_api_request_duration_seconds` metric buckets to correctly capture the latency metrics. Previously, the majority of the calls would fall in the "+Inf" bucket. ([#94873](https://github.com/kubernetes/kubernetes/pull/94873), [@marwanad](https://github.com/marwanad)) [SIG Cloud Provider and Instrumentation]

 *
 *  * http://host:80/
 *
 *  * http://host
 *
 * Both the unnecessary port specification (`:80`) and the absent trailing slash (`/`) cause URI to
 * bucket the two URLs separately. This harms URI's usefulness in collections. Any application that
 * stores information-per-URL will need to either canonicalize manually, or suffer unnecessary
 * redundancy for such URLs.
 *

  @J2ktIncompatible
  @GwtIncompatible // reflection
  @AndroidIncompatible // relies on assumptions about OpenJDK
  public void testNewHashMapWithExpectedSize_wontGrow() throws Exception {
    // before jdk7u40: creates one-bucket table
    // after  jdk7u40: creates empty table
    assertThat(bucketsOf(Maps.newHashMapWithExpectedSize(0))).isAtMost(1);

    for (int size = 1; size < 200; size++) {
      assertWontGrow(
          size,

- Fix the `cloudprovider_azure_api_request_duration_seconds` metric buckets to correctly capture the latency metrics. Previously, the majority of the calls would fall in the "+Inf" bucket. ([#94943](https://github.com/kubernetes/kubernetes/pull/94943), [@marwanad](https://github.com/marwanad)) [SIG Cloud Provider and Instrumentation]

    /** The key of the message: Secret Key */
    public static final String LABELS_storage_secret_key = "{labels.storage_secret_key}";

    /** The key of the message: Bucket */
    public static final String LABELS_storage_bucket = "{labels.storage_bucket}";

    /** The key of the message: Send Test Mail */
    public static final String LABELS_send_testmail = "{labels.send_testmail}";

- Metrics changes: Fix exposed buckets of `scheduler_volume_scheduling_duration_seconds_bucket` metric. ([#100720](https://github.com/kubernetes/kubernetes/pull/100720), [@dntosas](https://github.com/dntosas))

- Kube-proxy: delete stale conntrack UDP entries for loadbalancer ingress IP. ([#104151](https://github.com/kubernetes/kubernetes/pull/104151), [@aojea](https://github.com/aojea)) [SIG Network]
- Metrics changes: Fix exposed buckets of `scheduler_volume_scheduling_duration_seconds_bucket` metric ([#100720](https://github.com/kubernetes/kubernetes/pull/100720), [@dntosas](https://github.com/dntosas)) [SIG Apps, Instrumentation, Scheduling and Storage]

- Fix the `cloudprovider_azure_api_request_duration_seconds` metric buckets to correctly capture the latency metrics. Previously, the majority of the calls would fall in the "+Inf" bucket. ([#95376](https://github.com/kubernetes/kubernetes/pull/95376), [@marwanad](https://github.com/marwanad)) [SIG Cloud Provider and Instrumentation]