java - 使用 SpringFlux 的 webclient 重复 Mono

Question

情况如下:我发送了第一个请求，然后我以5秒的间隔重复发送了第二个请求。如果第二个请求收到成功响应，我会将其保存在数据库中并执行一些操作，如果它收到不成功(完整)响应，则应重复请求。当尝试次数大于 3 时，我需要停止重复第二个请求。我该如何做到这一点？

     return firstRequestSenderService.send(request)
        .flatMap(resp -> {
          AtomicInteger attempts = new AtomicInteger(0);
          String url = normalizeUrl(resp.getResult());
          return Mono.defer(() -> {
            log.info("Second request, attempt = {}, params = {}", attempts.get(), param);
            return secondRequestSenderService.send(param, url, attempts.getAndIncrement());
          })
              .filter(this::isCompleteResponse)
              // i try .filter(b -> attempts.get() > 2)
              .doOnNext(r -> log.info("Save report"))
              .map(secondResp -> dataSaver.saveReport(param, secondResp))
              .doOnNext(r -> log.info("Send request to another service"))
              .flatMap(r -> secondRequestSender.sendPdf(r)))
              .doOnNext(bytes -> dataSaver.saveAnotherReport(param, bytes))
              .repeatWhenEmpty(req -> Flux.interval(Duration.ofSeconds(5)));
// also try   .repeatWhenEmpty(3, req -> Flux.interval(Duration.ofSeconds(5)));
        })
        .then(Mono.empty());

Answer 1

检查reactor-addons你可以做类似的事情

    return firstRequestSenderService.send(request)
        .flatMap(resp -> {
          return secondRequestSenderService
                  .send(param, url, attempts.getAndIncrement())
                  .retryWhen(Retry.any()
                          .randomBackoff(Duration.ofMillis(100), Duration.ofMillis(5000))
                          .retryMax(3));

        });

你也可以使用reactor core retry utils RetryBackoffSpec来做到这一点当你使用最新的 spring webflux 时你会得到这一点

  RetryBackoffSpec retryBackoffSpec = Retry
        .fixedDelay(3, Duration.ofSeconds(5));
return firstRequestSenderService.send(request)
        .flatMap(resp -> {
                  return secondRequestSenderService
                          .send(param, url, attempts.getAndIncrement())
                          .retryWhen(retryBackoffSpec);
                });