助焊剂起死回生如何应对背压

Question

I need to handle back pressure in my flux, that receives as input a List of object. The size of the list varies from few hundreds to some hundreds of thousand elements. The actual code is:

Flux.fromIterable(alarms)
            .limitRate(parallelism)
            .parallel(parallelism)
            .runOn(Schedulers.elastic(), bufferSize)
            .doOnNext(reactiveHandleDataService::handleAlarm)
;

The parameter "limitRange" simply force to refuse a list over a certain size, thing that I don't want. I need all the data received to be given to reactiveHandleDataService, I can't lose a message.

How can I handle back pressure in this case? I did't find much example explaining well the problem, especially using an iterable as source.

I'm using Californium-SR3 as release of reactor and this is part of a spring boot application.

Answer 1

is handleAlarm job to have backpressure if cant support to many data to send request for new data after processing 1. if you cant have backPressure to handleAlarm you can add delay

Flux.fromIterable(alarms)
        .limitRate(parallelism)
        .delayElements(Duration.ofMillis(10))
        .doOnNext(reactiveHandleDataService::handleAlarm)
        .subscribeOn(Schedulars.elastic)

;

and if you want backpressure why you are running in parallel