/// 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.

Pour plus de détails, consultez la section [Déploiement](deployment/index.md){.internal-link target=_blank}.

///

## Benchmarks et vitesse

([#44744](https://github.com/kubernetes/kubernetes/pull/44744), [@yguo0905](https://github.com/yguo0905))

  * Kubernetes now shares a single PID namespace among all containers in a pod when running with docker >= 1.13.1. This means processes can now signal processes in other containers in a pod, but it also means that the `kubectl exec {pod} kill 1` pattern will cause the Pod to be restarted rather than a single container. ([#45236](https://github.com/kubernetes/kubernetes/pull/45236), [@verb](...

    /**
     * Get the value for the key 'job.max.crawler.processes'. <br>
     * The value is, e.g. 0 <br>
     * comment: Maximum number of crawler processes.
     * @return The value of found property. (NotNull: if not found, exception but basically no way)
     */
    String getJobMaxCrawlerProcesses();

    /**
     * Get the value for the key 'job.max.crawler.processes' as {@link Integer}. <br>
     * The value is, e.g. 0 <br>

for COS images on GCE has changed to `/home/kubernetes/flexvolume`, rather than `/etc/srv/kubernetes/kubelet-plugins/volume/exec`. Existing Flexvolume installations in clusters using COS images must be moved to the new directory, and installation processes must be updated with the new path. ([#58171](https://github.com/kubernetes/kubernetes/pull/58171), [@verult](https://github.com/verult))

* Default values differ between the Kubelet's componentconfig (config file) API and the Kubelet's...

    * Specifically the new controller watches the API server for scheduled pods. It processes each pod and ensures that any volumes that implement the volume Attacher interface are attached to the node their pod is scheduled to.

*...

- GCE/Windows: When the service cannot be stopped Stackdriver logging processes are now force killed ([#77378](https://github.com/kubernetes/kubernetes/pull/77378), [@yujuhong](https://github.com/yujuhong))

    * Each deployment method should bind users/groups to the ClusterRoles if they are using this feature.
* Adds AllowPrivilegeEscalation to control whether a process can gain more privileges than its parent process ([#47019](https://github.com/kubernetes/kubernetes/pull/47019), [@jessfraz](https://github.com/jessfraz))

- DRA: when asking for "All" devices on a node, Kubernetes <= 1.32 proceeded to schedule pods onto nodes with no devices by not allocating any devices for those pods. Kubernetes 1.33 changes that to only picking nodes which have at least one device. Users who want the "proceed with scheduling also without devices" semantic can use the upcoming prioritized list feature with one sub-request for "all" devices and a second...

- Added support for a new kubectl output format, `kyaml`. KYAML was a strict subset of YAML and should be accepted by any YAML processor. The formatting of KYAML was halfway between JSON and YAML.  Because it was more explicit than the default YAML style, it was less error-prone. ([#132942](https://github.com/kubernetes/kubernetes/pull/132942), [@thockin](https://github.com/thockin))...