* <p>The time taken to compute multiple quantiles on the same dataset using {@link Scale#indexes
 * indexes} is generally less than the total time taken to compute each of them separately, and
 * sometimes much less. For example, on a large enough dataset, computing the 90th and 99th
 * percentiles together takes about 55% as long as computing them separately.
 *
 * <p>When calling {@link ScaleAndIndex#compute} (in {@linkplain ScaleAndIndexes#compute either

* Kubernetes now supports transparent compression of API responses. Clients that send `Accept-Encoding: gzip` will now receive a GZIP compressed response body if the API call was larger than 128KB.  Go clients automatically request gzip-encoding by default and should see reduced transfer times for very large API requests.  Clients in other languages may need to make changes to benefit from compression. ([#77449](https://github.com/kubernetes/kubernetes/pull/77449), [@smarterclayton](https:...

- Fix the problem where couple nodes becoming NotReady at the same time could cause master instability or even complete outage in large enough clusters. ([#88962](https://github.com/kubernetes/kubernetes/pull/88962), [@mborsz](https://github.com/mborsz)) [SIG Apps]

* Audit logs might impact the API server performance and the latency of large request and response calls. The issue is observed under the following conditions: if `AdvancedAuditing` feature gate is enabled, which is the default case, if audit logging uses the log backend in JSON format, or if the audit policy records large API calls for requests or responses.

*   [beta] Flag `--chunk-size={SIZE}` is added to `kubectl get` to customize the number of results returned in large lists of resources. This reduces the perceived latency of managing large clusters because the server returns the first set of results to the client much more quickly. Pass 0 to disable this feature.([#53768](https://github.com/kubernetes/kubernetes/pull/53768),[ @smarterclayton](htt...

	erasureLockOwner string

	// List of endpoints provided on the command line.
	endpoints PoolEndpoints

	// String version of all the endpoints, an optimization
	// to avoid url.String() conversion taking CPU on
	// large disk setups.
	endpointStrings []string

	// Total number of sets and the number of disks per set.
	setCount, setDriveCount int
	defaultParityCount      int

	poolIndex int

	// Distribution algorithm of choice.

  {{.HelpName}} {{if .VisibleFlags}}[FLAGS] {{end}}DIR{1...64} DIR{65...128}

DIR:
  DIR points to a directory on a filesystem. When you want to combine
  multiple drives into a single large system, pass one directory per
  filesystem separated by space. You may also use a '...' convention
  to abbreviate the directory arguments. Remote directories in a
  distributed setup are encoded as HTTP(s) URIs.

 *       amounts of computation (creating a thread, calling a few methods, etc) needed to reach a
 *       timeout point. Similarly, a SMALL is always discriminable as larger than SHORT and smaller
 *       than MEDIUM. And so on. These constants are set to conservative values, but even so, if
 *       there is ever any doubt, they can all be increased in one spot to rerun tests on slower

# CL 134210043 archive/zip: add Writer.Flush, Brad Fitzpatrick <******@****.***>
pkg archive/zip, method (*Writer) Flush() error

# CL 97140043 compress/flate: add Reset() to allow reusing large buffers to compress multiple buffers, James Robinson <******@****.***>
pkg compress/flate, type Resetter interface { Reset }
pkg compress/flate, type Resetter interface, Reset(io.Reader, []uint8) error
pkg compress/zlib, type Resetter interface { Reset }

      }
    }
    Arrays.fill(contents, uniques, n, null);
    if (uniques < contents.length / 2) {
      // Deduplication eliminated many of the elements.  We don't want to retain an arbitrarily
      // large array relative to the number of elements, so we cap the ratio.
      contents = Arrays.copyOf(contents, uniques);
    }
    return new RegularImmutableSortedSet<E>(