Netflix/Hystrix v1.4.0

Version 1.4.0

on GitHub

This version adds HystrixObservableCommand and implements both it and HystrixCommand in terms of Observables.

A HystrixObservableCommand allows for fully non-blocking commands that can be composed as part of a larger Observable chain. See the wiki for more details on usage. Here's an example (using Java 8):

public class ObservableHttpCommand extends HystrixObsverableCommand<BackendResponse> {

@Override
protected Observable<BackendResponse> construct() {
    return httpClient.submit(HttpClientRequest.createGet("/mock.json?numItems=" + numItems))
        .flatMap((HttpClientResponse<ByteBuf> r) -> r.getContent()
        .map(b -> BackendResponse.fromJson(new ByteBufInputStream(b))));
    }

@Override
protected Observable<BackendResponse> resumeWithFallback() {
    return Observable.just(new BackendResponse(0, numItems, new String[] {}));
}}

Because an Observable represents a stream of data, your HystrixObservableCommand may now return a stream of data, and supply a stream of data as a fallback. The methods to do so are construct() and resumeWithFallback(), respectively. All other aspects of the Hystrix state machine work the same in a HystrixCommand. See this wiki page for a diagram of this state machine.

The public API of HystrixCommand is unchanged, though the internals have significantly changed. Some bugfixes are now possible that affect Hystrix semantics:

• Timeouts now apply to semaphore-isolated commands as well as thread-isolated commands. Before 1.4.x, semaphore-isolated commands could not timeout. They now have a timeout registered on another (HystrixTimer) thread, which triggers the timeout flow. If you use semaphore-isolated commands, they will now see timeouts. As all HystrixCommands have a default timeout, this potentially affects all semaphore-isolated commands.
• Timeouts now fire on HystrixCommand.queue(), even if the caller never calls get() on the resulting Future. Before 1.4.x, only calls to get() triggered the timeout mechanism to take effect.

You can see more in-depth examples of how the new functionality of Hystrix 1.4 is used at Netflix/ReactiveLab.

You can learn more about Observables and RxJava at Netflix/RxJava.

As this was a major refactoring of Hystrix internals, we (Netflix) have run a release candidate of Hystrix 1.4 in canaries and production over the last month to gain confidence.

• Pull 701 Fix Javadoc warnings
• Pull 700 Example code for publishing to Graphite

Artifacts: Maven Central / Bintray