@Test fun connectExceptionsAreReturnedAsSuppressed() {
    val proxySelector = RecordingProxySelector()
    proxySelector.proxies.add(Proxy(Proxy.Type.HTTP, TestUtil.UNREACHABLE_ADDRESS_IPV4))
    proxySelector.proxies.add(Proxy.NO_PROXY)
    server.close()

    client =
      client
        .newBuilder()
        .proxySelector(proxySelector)
        .readTimeout(Duration.ofMillis(100))

   | Unless otherwise specified (by system property or command-line switch), the first proxy
   | specification in this list marked as active will be used.
   |-->
  <proxies>
    <!-- proxy
     | Specification for one proxy, to be used in connecting to the network.
     |
    <proxy>
      <id>optional</id>
      <active>true</active>
      <protocol>http</protocol>

├── Dockerfile
└── requirements.txt
```

#### Derrière un proxy de terminaison TLS { #behind-a-tls-termination-proxy }

Si vous exécutez votre conteneur derrière un proxy de terminaison TLS (load balancer) comme Nginx ou Traefik, ajoutez l'option `--proxy-headers`, cela indiquera à Uvicorn (via la CLI FastAPI) de faire confiance aux en-têtes envoyés par ce proxy lui indiquant que l'application s'exécute derrière HTTPS, etc.

```Dockerfile

# Inclure WSGI - Flask, Django, autres { #including-wsgi-flask-django-others }

Vous pouvez monter des applications WSGI comme vous l'avez vu avec [Sous-applications - Montages](sub-applications.md), [Derrière un proxy](behind-a-proxy.md).

Pour cela, vous pouvez utiliser `WSGIMiddleware` et l'utiliser pour envelopper votre application WSGI, par exemple Flask, Django, etc.

## Utiliser `WSGIMiddleware` { #using-wsgimiddleware }

/// info

- 📫 Récupération de mot de passe par e-mail.
- ✅ Tests avec [Pytest](https://pytest.org).
- 📞 [Traefik](https://traefik.io) comme proxy inverse / répartiteur de charge.
- 🚢 Instructions de déploiement avec Docker Compose, y compris la configuration d'un proxy Traefik frontal pour gérer les certificats HTTPS automatiques.

    val listener1: EventListener =
      object : EventListener() {
        override fun connectStart(
          call: Call,
          inetSocketAddress: InetSocketAddress,
          proxy: Proxy,
        ) {
          try {
            // Wait for request2 to guarantee we make 2 separate connections to the server.
            latch1.await()
          } catch (e: InterruptedException) {

	public fun canceled (Lokhttp3/Call;)V
	public fun connectEnd (Lokhttp3/Call;Ljava/net/InetSocketAddress;Ljava/net/Proxy;Lokhttp3/Protocol;)V
	public fun connectFailed (Lokhttp3/Call;Ljava/net/InetSocketAddress;Ljava/net/Proxy;Lokhttp3/Protocol;Ljava/io/IOException;)V
	public fun connectStart (Lokhttp3/Call;Ljava/net/InetSocketAddress;Ljava/net/Proxy;)V
	public fun connectionAcquired (Lokhttp3/Call;Lokhttp3/Connection;)V

			Description: `staging dir for undelivered logger messages e.g. '/home/logger-events'`,
			Optional:    true,
			Type:        "string",
		},
		config.HelpKV{
			Key:         Proxy,
			Description: "proxy url endpoint e.g. http(s)://proxy",
			Optional:    true,
			Type:        "string",
		},
		config.HelpKV{
			Key:         MaxRetry,
			Description: `maximum retry count before we start dropping logged event(s)`,

 *    to use `H2_PRIOR_KNOWLEDGE` but the URL's scheme is `https:`.
 *  * Preemptive proxy authentication failed.
 *
 * Planning failures are not necessarily fatal. For example, even if we can't DNS lookup the first
 * proxy in a list, looking up a subsequent one may succeed.
 */
internal class FailedPlan(
  e: Throwable,
) : RoutePlanner.Plan {

# Incluyendo WSGI - Flask, Django, otros { #including-wsgi-flask-django-others }

Puedes montar aplicaciones WSGI como viste con [Sub Aplicaciones - Mounts](sub-applications.md), [Detrás de un Proxy](behind-a-proxy.md).

Para eso, puedes usar el `WSGIMiddleware` y usarlo para envolver tu aplicación WSGI, por ejemplo, Flask, Django, etc.

## Usando `WSGIMiddleware` { #using-wsgimiddleware }

/// info | Información