API. With that design, the QoS class for a pod only applied to CPU resources (such as `cpu_shares`). The Kubernetes cgroup manager uses `memory.limit_in_bytes` in v1 cgroups to limit the memory capacity for a container, and uses `oom_scores` to recommend an order for killing container processes if an out-of-memory event occurs. This implementation has shortcomings: for `Guaranteed` pods, memory can not be fully reserved, and the page cache is at risk of being recycled. For `Burstable` pods, overcommitting...

   * CacheBuilder#refreshAfterWrite}, or through a call to {@link LoadingCache#refresh}.
   *
   * <p>This implementation synchronously delegates to {@link #load}. It is recommended that it be
   * overridden with an asynchronous implementation when using {@link
   * CacheBuilder#refreshAfterWrite}.
   *
   * <p><b>Note:</b> <i>all exceptions thrown by this method will be logged and then swallowed</i>.

                        if ((previousRange != null) && (currentRange != null)) {
                            // TODO shouldn't need to double up on this work, only done for simplicity of handling
                            // recommended
                            // version but the restriction is identical
                            VersionRange newRange = previousRange.restrict(currentRange);

Il faut qu'il continue à exister pendant un certain temps car vos clients l'utilisent, mais vous voulez que la documentation mentionne clairement que ce paramètre est <abbr title="obsolète, recommandé de ne pas l'utiliser">déprécié</abbr>.

On utilise alors l'argument `deprecated=True` de `Query` :

{* ../../docs_src/query_params_str_validations/tutorial010.py hl[18] *}

    String getVersion();

    /**
     * {@return the project packaging}.
     * <p>
     * Note: unlike in legacy code, logical checks against string representing packaging (returned by this method)
     * are NOT recommended (code like {@code "pom".equals(project.getPackaging)} must be avoided). Use method
     * {@link #getArtifacts()} to gain access to POM or build artifact.
     *
     * @see #getArtifacts()
     */
    @Nonnull

quelque chose qu'un framework comme Starlette (ou **FastAPI**) fournirait par-dessus.

C'est le serveur recommandé pour Starlette et **FastAPI**.

/// check | **FastAPI** le recommande comme

Le serveur web principal pour exécuter les applications **FastAPI**.

Vous pouvez le combiner avec Gunicorn, pour avoir un serveur multi-processus asynchrone.

 * used weak references, the entries are reconstructed as they were, but it's not unlikely they'll
 * be quickly garbage-collected before they are ever accessed.
 *
 * <p>{@code new MapMaker().weakKeys().makeMap()} is a recommended replacement for {@link
 * java.util.WeakHashMap}, but note that it compares keys using object identity whereas {@code
 * WeakHashMap} uses {@link Object#equals}.
 *
 * @author Bob Lee
 * @author Charles Fry

or more targets are participating in active-active replication, the replication configuration for replica metadata sync, delete marker replication and delete replication should match to avoid inconsistent picture between the clusters. It is not recommended to turn on asymmetric replication - for e.g. if three sites A,B,C are participating in replication, it would be better to avoid replication setups like A -> [B, C], B -> A. In this particular example, an object uploaded to A will be replicated to...

### Action required

  * Docker v1.9.1 is officially recommended. Docker v1.8.3 and Docker v1.10 are
supported. If you are using an older release of Docker, please upgrade. Known
issues with Docker 1.9.1 can be found below.

        }

        /**
         * Filters the types of paths to include in the result.
         * The result will contain only the paths of types for which the predicate returned {@code true}.
         * It is recommended to apply a filter for retaining only the types of paths of interest,
         * because it can resolve ambiguities when a path could be of many types.
         *