### Concurrence et Burgers

L'idée de code **asynchrone** décrite ci-dessus est parfois aussi appelée **"concurrence"**. Ce qui est différent du **"parallélisme"**.

La **concurrence** et le **parallélisme** sont tous deux liés à l'idée de "différentes choses arrivant plus ou moins au même moment".

Mais les détails entre la **concurrence** et le **parallélisme** diffèrent sur de nombreux points.

- Added a new metric `apiserver_flowcontrol_request_concurrency_in_use` that shows the number of
  seats (concurrency) occupied by the currently executing requests in the API Priority and Fairness system. ([#102795](https://github.com/kubernetes/kubernetes/pull/102795), [@tkashem](https://github.com/tkashem))

  // priority level that may be exeucting at a given time.  ACS must
  // be a positive number. The server's concurrency limit (SCL) is
  // divided among the concurrency-controlled priority levels in
  // proportion to their assured concurrency shares. This produces
  // the assured concurrency value (ACV) --- the number of requests
  // that may be executing at a time --- for each such priority
  // level:
  //

- Watch requests that are delegated to aggregated API servers no longer reserve concurrency units (seats) in the API Priority and Fairness dispatcher for their entire duration. ([#105511](https://github.com/kubernetes/kubernetes/pull/105511), [@benluddy](https://github.com/benluddy))

*   `tf.config.experimental_list_devices` has been removed, please use
    `tf.config.list_logical_devices`.

## Bug Fixes and Other Changes

*   `tf.data`
    *   Fixes concurrency issue with `tf.data.experimental.parallel_interleave`
        with `sloppy=True`.
    *   Add `tf.data.experimental.dense_to_ragged_batch()`.
    *   Extend `tf.data` parsing ops to support `RaggedTensors`.

- Update kube-apiserver's priority & fairness work estimator such that 'max seats' is MIN(0.15 x nominalCL, nominalCL / handSize)
  
  This fixes a bug where clients with requests using hand size x max seats greater than the nominal concurrency limit can starve other requests in the same priority level. ([#118601](https://github.com/kubernetes/kubernetes/pull/118601), [@andrewsykim](https://github.com/andrewsykim)) [SIG API Machinery]

  // priority level that may be exeucting at a given time.  ACS must
  // be a positive number. The server's concurrency limit (SCL) is
  // divided among the concurrency-controlled priority levels in
  // proportion to their assured concurrency shares. This produces
  // the assured concurrency value (ACV) --- the number of requests
  // that may be executing at a time --- for each such priority
  // level:
  //

  // priority level that may be exeucting at a given time.  ACS must
  // be a positive number. The server's concurrency limit (SCL) is
  // divided among the concurrency-controlled priority levels in
  // proportion to their assured concurrency shares. This produces
  // the assured concurrency value (ACV) --- the number of requests
  // that may be executing at a time --- for each such priority
  // level:
  //

import okhttp3.Dispatcher
import okhttp3.internal.http2.Http2Connection
import okhttp3.internal.http2.Http2Stream
import okhttp3.internal.http2.Http2Writer

/**
 * Centralisation of central locks according to docs/contribute/concurrency.md
 */
internal object Locks {
  inline fun <T> Dispatcher.withLock(action: () -> T): T {
    contract { callsInPlace(action, InvocationKind.EXACTLY_ONCE) }
    return lock.withLock(action)
  }

  // borrowing seats from this level.
  // The server's concurrency limit (ServerCL) is divided among the
  // Limited priority levels in proportion to their NCS values:
  //
  // NominalCL(i)  = ceil( ServerCL * NCS(i) / sum_ncs )
  // sum_ncs = sum[limited priority level k] NCS(k)
  //
  // Bigger numbers mean a larger nominal concurrency limit,
  // at the expense of every other Limited priority level.