Class ReactorClientHttpRequestFactory

java.lang.Object
org.springframework.http.client.ReactorClientHttpRequestFactory
All Implemented Interfaces:
Lifecycle, Phased, SmartLifecycle, ClientHttpRequestFactory
public class ReactorClientHttpRequestFactory extends Object implements ClientHttpRequestFactory, SmartLifecycle
Reactor-Netty implementation of ClientHttpRequestFactory.

This class implements SmartLifecycle and can be optionally declared as a Spring-managed bean in order to support JVM Checkpoint Restore.

Since:
6.2
Author:
Arjen Poutsma, Juergen Hoeller, Sebastien Deleuze, Brian Clozel

  • Constructor Details

    • ReactorClientHttpRequestFactory

      public ReactorClientHttpRequestFactory()
      Constructor with default client, created via HttpClient.create(), and with compression and proxyWithSystemProperties enabled.

    • ReactorClientHttpRequestFactory

      public ReactorClientHttpRequestFactory(reactor.netty.http.client.HttpClient client)
      Constructor with a given HttpClient instance.
      Parameters:
      client - the client to use

    • ReactorClientHttpRequestFactory

      public ReactorClientHttpRequestFactory(ReactorResourceFactory resourceFactory, Function<reactor.netty.http.client.HttpClient, reactor.netty.http.client.HttpClient> mapper)
      Constructor with externally managed Reactor Netty resources, including LoopResources for event loop threads, and ConnectionProvider for connection pooling.

      Generally, it is recommended to share resources for event loop concurrency. This can be achieved either by participating in the JVM-wide, global resources held in HttpResources, or by using a specific, shared set of resources through a ReactorResourceFactory bean. The latter can ensure that resources are shut down when the Spring ApplicationContext is stopped/closed and restarted again (e.g. JVM checkpoint restore).

      Parameters:
      resourceFactory - the resource factory to get resources from
      mapper - for further initialization of the client

  • Method Details

    • setExecutor

      public void setExecutor(Executor executor)
      Set the Executor to use for performing blocking I/O operations.

      If no executor is provided, the request will use an elastic scheduler.

      Parameters:
      executor - the executor to use.
      Since:
      6.2.13

    • setConnectTimeout

      public void setConnectTimeout(int connectTimeout)
      Set the connect timeout on the underlying client, in milliseconds.

      Effectively, a shortcut for httpClient.option(CONNECT_TIMEOUT_MILLIS, connectTimeout).

      A timeout value of 0 specifies an infinite timeout.

      Negative values are not permitted.

      Default is 30 seconds, as defined by Reactor Netty.

      Parameters:
      connectTimeout - the connect timeout in milliseconds
      See Also:

    • setConnectTimeout

      public void setConnectTimeout(Duration connectTimeout)
      Set the connect timeout on the underlying client, as a Duration.

      Effectively, a shortcut for httpClient.option(CONNECT_TIMEOUT_MILLIS, connectTimeout.toMillis()).

      A timeout value of 0 specifies an infinite timeout.

      Negative values and values less than one millisecond (other than 0) are not permitted, and values exceeding Integer.MAX_VALUE milliseconds are limited to Integer.MAX_VALUE milliseconds.

      Default is 30 seconds, as defined by Reactor Netty.

      Parameters:
      connectTimeout - the connect timeout as a Duration
      See Also:

    • setReadTimeout

      public void setReadTimeout(Duration readTimeout)
      Set the read timeout on the underlying client, as a Duration.

      Effectively, a shortcut for HttpClient.responseTimeout(Duration).

      A timeout value of 0 specifies an infinite timeout.

      Negative values and values less than one millisecond (other than 0) are not permitted.

      Default is 10 seconds, unless a pre-configured HttpClient is supplied.

      Parameters:
      readTimeout - the read timeout as a Duration
      See Also:

    • setReadTimeout

      public void setReadTimeout(long readTimeout)
      Set the read timeout on the underlying client, in milliseconds.

      Effectively, a shortcut for httpClient.responseTimeout(Duration.ofMillis(readTimeout)).

      A timeout value of 0 specifies an infinite timeout.

      Negative values are not permitted.

      Default is 10 seconds, unless a pre-configured HttpClient is supplied.

      Parameters:
      readTimeout - the read timeout in milliseconds
      See Also:

    • createRequest

      public ClientHttpRequest createRequest(URI uri, HttpMethod httpMethod) throws IOException
      Description copied from interface: ClientHttpRequestFactory
      Create a new ClientHttpRequest for the specified URI and HTTP method.

      The returned request can be written to, and then executed by calling ClientHttpRequest.execute().

      Specified by:
      createRequest in interface ClientHttpRequestFactory
      Parameters:
      uri - the URI to create a request for
      httpMethod - the HTTP method to execute
      Returns:
      the created request
      Throws:
      IOException - in case of I/O errors

    • start

      public void start()
      Description copied from interface: Lifecycle
      Start this component.

      Should not throw an exception if the component is already running.

      In the case of a container, this will propagate a hard start signal to all components that apply, even to non-auto-startup components.

      Specified by:
      start in interface Lifecycle
      See Also:

    • stop

      public void stop()
      Description copied from interface: Lifecycle
      Stop this component, typically in a synchronous fashion, such that the component is fully stopped upon return of this method. Consider implementing SmartLifecycle and its stop(Runnable) variant when asynchronous stop behavior is necessary.

      Note that this stop notification is not guaranteed to come before destruction: On regular shutdown, Lifecycle beans will first receive a stop notification before the general destruction callbacks are being propagated; however, on hot refresh during a context's lifetime or on aborted refresh attempts, a given bean's destroy method will be called without any consideration of stop signals upfront.

      Should not throw an exception if the component is not running (not started yet).

      In the case of a container, this will propagate the stop signal to all components that apply.

      Specified by:
      stop in interface Lifecycle
      See Also:

    • isRunning

      public boolean isRunning()
      Description copied from interface: Lifecycle
      Check whether this component is currently running.

      In the case of a container, this will return true only if all components that apply are currently running.

      Specified by:
      isRunning in interface Lifecycle
      Returns:
      whether the component is currently running

    • getPhase

      public int getPhase()
      Description copied from interface: SmartLifecycle
      Return the phase that this lifecycle object is supposed to run in.

      The default implementation returns SmartLifecycle.DEFAULT_PHASE in order to let stop() callbacks execute before regular Lifecycle implementations.

      Specified by:
      getPhase in interface Phased
      Specified by:
      getPhase in interface SmartLifecycle
      See Also: