Develop Reference

Deal efficiently with a lot of discarded objects

I have the following reactive stream where logs is Flux<String>:
logs.bufferTimeout(50, Duration.ofSeconds(20))
.doOnNext(logs -> sendLogs(logs))
.doOnDiscard(String.class, log -> sendLogs(List.of(log)));
The purpose of the stream is to send logs to another microservice for long-term storage. As you can see, the logs are buffered and sent in batches of 50. The intent is to limit the number of requests made to the second microservice.
For reasons that are not relevant to this question, the logs stream can be canceled at any time. Whenever this happens, I want all log messages that are still in the buffer to also be sent to the second microservice. The doOnDiscard operator I've added does just that, but perhaps not in the most performant/safe way. Let's say that there are 49 logs in the buffer at the time of cancellation. The lambda in doOnDiscard will then be called 49 times (once for each discarded log message) and this will result in 49 requests to the second microservice is a short amount of time.
You can reproduce this behavior by executing the following code:
public static void main(String[] args) throws InterruptedException {
Disposable disposable = Flux.interval(Duration.ofSeconds(1))
.bufferTimeout(10, Duration.ofSeconds(5))
.doOnNext(System.out::println)
.doOnDiscard(Long.class, i -> System.out.println("Discarded: " + i))
.subscribe();
Thread.sleep(8000);
disposable.dispose();
Thread.sleep(2000);
}
Is there any way to buffer the discarded elements and process them together? I've tried the following:
public static void main(String[] args) throws InterruptedException {
List<Long> discardedElements = new ArrayList<>();
Disposable disposable = Flux.interval(Duration.ofSeconds(1))
.bufferTimeout(10, Duration.ofSeconds(5))
.doOnNext(System.out::println)
.doOnDiscard(Long.class, discardedElements::add)
.doFinally(signal -> System.out.println("Discarded elements: " + discardedElements))
.subscribe();
Thread.sleep(8000);
disposable.dispose();
Thread.sleep(2000);
}
It does what I want, but I feel like using a variable outside of the stream for temporary storage is not the "cleanest" solution. Is there a better one? Is the pattern of using an external variable a "recommended" one?

Reactor - how to retry on hot flux without dropping elements?

I have an infinite hot flux of data. I am about to engage in carrying out an operation on each element in the stream, each of which returns a Mono which will complete (one way or another) after some finite time.
There is the possibility of an error being thrown from these operations. If so, I want to resubscribe to the hot flux without missing anything, retrying elements that were in the middle of being processed when the error was thrown (i.e. anything that did not complete successfully).
What do I do here? I can tolerate repeated operations on the same elements, but not losing elements entirely from the stream.
I've attempted to use a ReplayProcessor to handle this, but I can't see a way of making it work without repeating a lot of operations that might well have succeeded (using a very conservative timeout), or losing elements due to new elements overriding old ones in the buffer (as below).
Test case:
#Test
public void fluxTest() {
List<String> strings = new ArrayList<>();
strings.add("one");
strings.add("two");
strings.add("three");
strings.add("four");
ConnectableFlux<String> flux = Flux.fromIterable(strings).publish();
//Goes boom after three uses of its method, otherwise
//returns a mono. completing after a little time
DangerousClass dangerousClass = new DangerousClass(3);
ReplayProcessor<String> replay = ReplayProcessor.create(3);
flux.subscribe(replay);
replay.flatMap(dangerousClass::doThis)
.retry(1)
.doOnNext(s -> LOG.info("Completed {}", s))
.subscribe();
flux.connect();
flux.blockLast();
}
public class DangerousClass {
Logger LOG = LoggerFactory.getLogger(DangerousClass.class);
private int boomCount;
private AtomicInteger count;
public DangerousClass(int boomCount) {
this.boomCount = boomCount;
this.count = new AtomicInteger(0);
}
public Mono<String> doThis(String s) {
return Mono.fromSupplier(() -> {
LOG.info("doing dangerous {}", s);
if (count.getAndIncrement() == boomCount) {
LOG.error("Throwing exception from {}", s);
throw new RuntimeException("Boom!");
}
return s;
}).delayElement(Duration.ofMillis(600));
}
}
This prints:
doing dangerous one
doing dangerous two
doing dangerous three
doing dangerous four
Throwing exception from four
doing dangerous two
doing dangerous three
doing dangerous four
Completed four
Completed two
Completed three
One is never completed.

The error (at least in the above example) can only occur in the flatMap(dangerousClass::doThis) call - so resubscribing to the root Flux and replaying elements when this one flatMap() call has failed seems a bit odd, and (probably) isn't what you want to do.
Instead, I'd recommend ditching the ReplayProcessor and just calling retry on the inner flatMap() call instead, so you end up with something like:
ConnectableFlux<String> flux = Flux.range(1, 10).map(n -> "Entry " + n).publish();
DangerousClass dangerousClass = new DangerousClass(3);
flux.flatMap(x -> dangerousClass.doThis(x).retry(1))
.doOnNext(s -> System.out.println("Completed " + s))
.subscribe();
flux.connect();
This will give you something like the following, with all entries completed and no retries:
doing dangerous Entry 1
doing dangerous Entry 2
doing dangerous Entry 3
doing dangerous Entry 4
Throwing exception from Entry 4
doing dangerous Entry 4
Completed Entry 2
Completed Entry 1
Completed Entry 3
Completed Entry 4

Limiting rate of requests with Reactor

I'm using project reactor to load data from a web service using rest. This is done in parallel with multiple threads. I'm starting to hit rate limits on the web service, so I would like to send at most 10 requests per second to avoid getting these errors. How would I do that using reactor?
Using zipWith(Mono.delayMillis(100))? Or is there some better way?
Thank you

You can use delayElements instead of the whole zipwith.

One could use Flux.delayElements to process a 10 requests batch at every 1s; be aware though that if the processing takes longer than 1s the next batch will still be started in parallel hence being processed together with the previous one (and potentially many other previous ones)!
That's why I propose another solution where a 10 requests batch is still processed at every 1s but, if its processing takes longer than 1s, the next batch will fail (see overflow IllegalStateException); one could deal with that failure such that to continue the overall processing but I won't show that here because I want to keep the example simple; see onErrorResume useful to handle overflow IllegalStateException.
The code below will do a GET on https://www.google.com/ at a rate of 10 requests per second. You'll have to do additional changes in order to support the situation where your server is not able to process in 1s all your 10 requests; you could just skip sending requests when those asked at previous second are still processed by your server.
#Test
void parallelHttpRequests() {
// this is just for limiting the test running period otherwise you don't need it
int COUNT = 2;
// use whatever (blocking) http client you desire;
// when using e.g. WebClient (Spring, non blocking client)
// the example will slightly change for no longer use
// subscribeOn(Schedulers.elastic())
RestTemplate client = new RestTemplate();
// exit, lock, condition are provided to allow one to run
// all this code in a #Test, otherwise they won't be needed
var exit = new AtomicBoolean(false);
var lock = new ReentrantLock();
var condition = lock.newCondition();
MessageFormat message = new MessageFormat("#batch: {0}, #req: {1}, resultLength: {2}");
Flux.interval(Duration.ofSeconds(1L))
.take(COUNT) // this is just for limiting the test running period otherwise you don't need it
.doOnNext(batch -> debug("#batch", batch)) // just for debugging
.flatMap(batch -> Flux.range(1, 10) // 10 requests per 1 second
.flatMap(i -> Mono.fromSupplier(() ->
client.getForEntity("https://www.google.com/", String.class).getBody()) // your request goes here (1 of 10)
.map(s -> message.format(new Object[]{batch, i, s.length()})) // here the request's result will be the output of message.format(...)
.doOnSubscribe(s -> debug("doOnSubscribe: #batch = " + batch + ", i = " + i)) // just for debugging
.subscribeOn(Schedulers.elastic()) // one I/O thread per request
)
)
// consider using onErrorResume to handle overflow IllegalStateException
.subscribe(
s -> debug("received", s) // do something with the above request's result
e -> {
// pay special attention to overflow IllegalStateException
debug("error", e.getMessage());
signalAll(exit, condition, lock);
},
() -> {
debug("done");
signalAll(exit, condition, lock);
}
);
await(exit, condition, lock);
}
// you won't need the "await" and "signalAll" methods below which
// I created only to be easier for one to run this in a test class
private void await(AtomicBoolean exit, Condition condition, Lock lock) {
lock.lock();
while (!exit.get()) {
try {
condition.await();
} catch (InterruptedException e) {
// maybe spurious wakeup
e.printStackTrace();
}
}
lock.unlock();
debug("exit");
}
private void signalAll(AtomicBoolean exit, Condition condition, Lock lock) {
exit.set(true);
try {
lock.lock();
condition.signalAll();
} finally {
lock.unlock();
}
}

Generate infinite sequence of Natural numbers using RxJava

I am trying to write a simple program using RxJava to generate an infinite sequence of natural numbers. So, far I have found two ways to generate sequence of numbers using Observable.timer() and Observable.interval(). I am not sure if these functions are the right way to approach this problem. I was expecting a simple function like one we have in Java 8 to generate infinite natural numbers.
IntStream.iterate(1, value -> value +1).forEach(System.out::println);
I tried using IntStream with Observable but that does not work correctly. It sends infinite stream of numbers only to first subscriber. How can I correctly generate infinite natural number sequence?
import rx.Observable;
import rx.functions.Action1;
import java.util.stream.IntStream;
public class NaturalNumbers {
public static void main(String[] args) {
Observable<Integer> naturalNumbers = Observable.<Integer>create(subscriber -> {
IntStream stream = IntStream.iterate(1, val -> val + 1);
stream.forEach(naturalNumber -> subscriber.onNext(naturalNumber));
});
Action1<Integer> first = naturalNumber -> System.out.println("First got " + naturalNumber);
Action1<Integer> second = naturalNumber -> System.out.println("Second got " + naturalNumber);
Action1<Integer> third = naturalNumber -> System.out.println("Third got " + naturalNumber);
naturalNumbers.subscribe(first);
naturalNumbers.subscribe(second);
naturalNumbers.subscribe(third);
}
}

The problem is that the on naturalNumbers.subscribe(first);, the OnSubscribe you implemented is being called and you are doing a forEach over an infinite stream, hence why your program never terminates.
One way you could deal with it is to asynchronously subscribe them on a different thread. To easily see the results I had to introduce a sleep into the Stream processing:
Observable<Integer> naturalNumbers = Observable.<Integer>create(subscriber -> {
IntStream stream = IntStream.iterate(1, i -> i + 1);
stream.peek(i -> {
try {
// Added to visibly see printing
Thread.sleep(50);
} catch (InterruptedException e) {
}
}).forEach(subscriber::onNext);
});
final Subscription subscribe1 = naturalNumbers
.subscribeOn(Schedulers.newThread())
.subscribe(first);
final Subscription subscribe2 = naturalNumbers
.subscribeOn(Schedulers.newThread())
.subscribe(second);
final Subscription subscribe3 = naturalNumbers
.subscribeOn(Schedulers.newThread())
.subscribe(third);
Thread.sleep(1000);
System.out.println("Unsubscribing");
subscribe1.unsubscribe();
subscribe2.unsubscribe();
subscribe3.unsubscribe();
Thread.sleep(1000);
System.out.println("Stopping");

Observable.Generate is exactly the operator to solve this class of problem reactively. I also assume this is a pedagogical example, since using an iterable for this is probably better anyway.
Your code produces the whole stream on the subscriber's thread. Since it is an infinite stream the subscribe call will never complete. Aside from that obvious problem, unsubscribing is also going to be problematic since you aren't checking for it in your loop.
You want to use a scheduler to solve this problem - certainly do not use subscribeOn since that would burden all observers. Schedule the delivery of each number to onNext - and as a last step in each scheduled action, schedule the next one.
Essentially this is what Observable.generate gives you - each iteration is scheduled on the provided scheduler (which defaults to one that introduces concurrency if you don't specify it). Scheduler operations can be cancelled and avoid thread starvation.
Rx.NET solves it like this (actually there is an async/await model that's better, but not available in Java afaik):
static IObservable<int> Range(int start, int count, IScheduler scheduler)
{
return Observable.Create<int>(observer =>
{
return scheduler.Schedule(0, (i, self) =>
{
if (i < count)
{
Console.WriteLine("Iteration {0}", i);
observer.OnNext(start + i);
self(i + 1);
}
else
{
observer.OnCompleted();
}
});
});
}
Two things to note here:
The call to Schedule returns a subscription handle that is passed back to the observer
The Schedule is recursive - the self parameter is a reference to the scheduler used to call the next iteration. This allows for unsubscription to cancel the operation.
Not sure how this looks in RxJava, but the idea should be the same. Again, Observable.generate will probably be simpler for you as it was designed to take care of this scenario.

When creating infinite sequencies care should be taken to:
subscribe and observe on different threads; otherwise you will only serve single subscriber
stop generating values as soon as subscription terminates; otherwise runaway loops will eat your CPU
The first issue is solved by using subscribeOn(), observeOn() and various schedulers.
The second issue is best solved by using library provided methods Observable.generate() or Observable.fromIterable(). They do proper checking.
Check this:
Observable<Integer> naturalNumbers =
Observable.<Integer, Integer>generate(() -> 1, (s, g) -> {
logger.info("generating {}", s);
g.onNext(s);
return s + 1;
}).subscribeOn(Schedulers.newThread());
Disposable sub1 = naturalNumbers
.subscribe(v -> logger.info("1 got {}", v));
Disposable sub2 = naturalNumbers
.subscribe(v -> logger.info("2 got {}", v));
Disposable sub3 = naturalNumbers
.subscribe(v -> logger.info("3 got {}", v));
Thread.sleep(100);
logger.info("unsubscribing...");
sub1.dispose();
sub2.dispose();
sub3.dispose();
Thread.sleep(1000);
logger.info("done");

Java 8 parallel streams don't appear to actually be working in parallel

I'm trying to use Java 8's parallelStream() to execute several long-running requests (eg web requests) in parallel. Simplified example:
List<Supplier<Result>> myFunctions = Arrays.asList(() -> doWebRequest(), ...)
List<Result> results = myFunctions.parallelStream().map(function -> function.get()).collect(...
So if there are two functions that block for 2 and 3 seconds respectively, I'd expect to get the result after 3 seconds. However, it really takes 5 seconds - ie it seems the functions are being executed in sequence and not in parallel. Am I doing something wrong?
edit: This is an example. The time taken is ~4000 milliseconds when I want it to be ~2000.
long start = System.currentTimeMillis();
Map<String, Supplier<String>> input = new HashMap<String, Supplier<String>>();
input.put("1", () -> {
try {
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
}
return "a";
});
input.put("2", () -> {
try {
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
}
return "b";
});
Map<String, String> results = input.keySet().parallelStream().collect(Collectors.toConcurrentMap(
key -> key,
key -> {
return input.get(key).get();
}));
System.out.println("Time: " + (System.currentTimeMillis() - start));
}
Doesn't make any difference if I iterate over the entrySet() instead of the keySet()
edit: changing the parallel part to the following also does not help:
Map<String, String> results = input.entrySet().parallelStream().map(entry -> {
return new ImmutablePair<String, String>(entry.getKey(), entry.getValue().get());
}).collect(Collectors.toConcurrentMap(Pair::getLeft, Pair::getRight));

When executing in parallel, there is overhead of decomposing the input set, creating tasks to represent the different portions of the calculation, distributing the actions across threads, waiting for results, combining results, etc. This is over and above the work of actually solving the problem. If a parallel framework were to always decompose problems down to a granularity of one element, for most problems, these overheads would overwhelm the actual computation and parallelism would result in a slower execution. So parallel frameworks have some latitude to decide how finely to decompose the input, and that's what's happening here.
In your case, your input set is simply too small to be decomposed. So the library chooses to execute sequentially.
Try this on your four-core system: compare
IntStream.range(0, 100_000).sum()
vs
IntStream.range(0, 100_000).parallel().sum()
Here, you're giving it enough input that it will be confident it can win through parallel execution. If you measure with a responsible measurement methodology (say, the JMH microbenchmark harness), you'll probably see an almost-linear speedup between these two examples.