If the response issues an authorization challenge, OkHttp will ask the [`Authenticator`](https://square.github.io/okhttp/5.x/okhttp/okhttp3/-authenticator/) (if one is configured) to satisfy the challenge. If the authenticator supplies a credential, the request is retried with that credential included.

## Retrying Requests

 4. If it's a new route, it connects by building either a direct socket connection, a TLS tunnel (for HTTPS over an HTTP proxy), or a direct TLS connection. It does TLS handshakes as necessary. This step may be retried for tunnel challenges and TLS handshake failures.
 5. It sends the HTTP request and reads the response.

    // If we do have a route selector, use its routes.
    return localRouteSelector.hasNext()
  }

  /**
   * Return the route from [connection] if it should be retried, even if the connection itself is
   * unhealthy. The biggest gotcha here is that we shouldn't reuse routes from coalesced
   * connections.
   */
  private fun retryRoute(connection: RealConnection): Route? =

        sslSocket.closeQuietly()
      }
    }
  }

  /**
   * To make an HTTPS connection over an HTTP proxy, send an unencrypted CONNECT request to create
   * the proxy connection. This may need to be retried if the proxy requires authorization.
   */
  @Throws(IOException::class)
  private fun createTunnel(): Request? {
    var nextRequest = tunnelRequest!!

        assertEquals(3, engine.evaluate(script, params2));

        engine.close();
    }

    // ===== Compilation Failure Tests =====

    /**
     * Test that syntax errors are not cached and can be retried
     */
    @Test
    public void test_evaluate_syntaxErrorNotCached() {
        final GroovyEngine engine = new GroovyEngine();
        final Map<String, Object> params = new HashMap<>();

    taskFaker.runNextTask()

    // Confirm the partial snapshot is not returned.
    assertThat(cache["a"]).isNull()

    // Confirm we prevent edits after a trim failure.
    assertThat(cache.edit("a")).isNull()

    // Confirm the partial snapshot is not returned after a successful trim.
    filesystem.setFaultyDelete(cacheDir / "a.1", false)
    taskFaker.runNextTask()

- Reflecting the...

    callback.await(request.url).assertHandshake()
  }

  @Test
  fun recoverWhenRetryOnConnectionFailureIsTrue() {
    // Set to 2 because the seeding request will count down before the retried request does.
    val requestFinished = CountDownLatch(2)
    val dispatcher: QueueDispatcher =
      object : QueueDispatcher() {
        override fun dispatch(request: RecordedRequest): MockResponse {

- Fixed a bug in `kube-scheduler` where pending pod preemption caused preemptor pods to be retried more frequently. ([#134245](https://github.com/kubernetes/kubernetes/pull/134245), [@macsko](https://github.com/macsko)) [SIG Scheduling and Testing]

  }

  private fun testRequestBodySurvivesRetries(transferKind: TransferKind) {
    server.enqueue(MockResponse(body = "abc"))

    // Use a misconfigured proxy to guarantee that the request is retried.
    client =
      client
        .newBuilder()
        .proxySelector(
          FakeProxySelector()
            .addProxy(server2.proxyAddress)
            .addProxy(Proxy.NO_PROXY),
        ).build()