*
   * The deadline to respond to this ping attempts to limit the cost of being wrong. If it is too
   * long, streams created while we await the pong will reuse broken connections and inevitably
   * fail. If it is too short, slow connections will be marked as failed and extra TCP and TLS
   * handshakes will be required.
   *

      factory.setAccessible(true);
    }
    // Sorts methods/constructors with the least number of parameters first since it's likely easier
    // to fill dummy parameter values for them. Ties are broken by name then by the string form of
    // the parameter list.
    return BY_NUMBER_OF_PARAMETERS
        .compound(BY_METHOD_NAME)
        .compound(BY_PARAMETERS)
        .immutableSortedCopy(factories);
  }

            // the tconHostName is from the DC referral, that referral must be resolved
            // before following deeper ones. Otherwise e.g. samba will return a broken
            // referral.
            synchronized (trans) {
                try {
                    // ensure connected
                    trans.ensureConnected();

- HTTP/2 connection health check is enabled by default in all Kubernetes clients to fix persistently broken connections (https://github.com/kubernetes/client-go/issues/374). If needed, users can tune the feature via the HTTP2_READ_IDLE_TIMEOUT_SECONDS and HTTP2_PING_TIMEOUT_SECONDS environment variables. The feature is disabled if HTTP2_READ_IDLE_TIMEOUT_SECONDS...

                    log.trace(Hexdump.toHexString(buffer, 4, n));

                }
                /*
                 * For some reason this can sometimes get broken up into another
                 * "NBSS Continuation Message" frame according to WireShark
                 */

                this.out.write(buffer, 0, 4 + n);
                this.out.flush();
            }

    server.enqueue(mockResponse)
    val responseAfter = MockResponse(body = "This comes after a busted connection")
    server.enqueue(responseAfter)
    server.enqueue(responseAfter) // Enqueue 2x because the broken connection may be reused.
    val response1 = getResponse(newRequest("/a"))
    response1.body
      .source()
      .timeout()
      .timeout(100, TimeUnit.MILLISECONDS)

    LoadingCache<String, String> cache =
        builder.build(
            new CacheLoader<String, String>() {
              @Override
              @SuppressWarnings("CacheLoaderNull") // test of broken user implementation
              public String load(String key) throws InterruptedException {
                callCount.incrementAndGet();
                startSignal.await();
                return null;

* Fix `kubectl logs -f` for windows server containers. ([#81747](https://github.com/kubernetes/kubernetes/pull/81747), [@Random-Liu](https://github.com/Random-Liu))
* Fix VMSS LoadBalancer backend pools so that the network won't be broken when instances are upgraded to latest model ([#81411](https://github.com/kubernetes/kubernetes/pull/81411), [@nilo19](https://github.com/nilo19))

    LoadingCache<String, String> cache =
        builder.build(
            new CacheLoader<String, String>() {
              @Override
              @SuppressWarnings("CacheLoaderNull") // test of broken user implementation
              public String load(String key) throws InterruptedException {
                callCount.incrementAndGet();
                startSignal.await();
                return null;

- Fixed an issue with setting `distinctAttribute=nil` when the `DRAConsumableCapacity` feature gate is disabled. ([#134962](https://github.com/kubernetes/kubernetes/pull/134962), [@sunya-ch](https://github.com/sunya-ch))
- Fixed broken shell completion for API resources. ([#133771](https://github.com/kubernetes/kubernetes/pull/133771), [@marckhouzam](https://github.com/marckhouzam))