1.非響應式代碼

    // 使用reactive 耗時：6.7s  使用mysql-2線程池 redis-2線程池
   // 使用常規(guī)的線程池 耗時：13s
    @PostMapping(value = "testWrite2Redis")
    public Result<Long> testRedis() throws Exception {
        final ExecutorService executorService = Executors.newFixedThreadPool(32);
        long begin = System.currentTimeMillis();
        ValueOperations<String, String> valOpt = redisTemplate.opsForValue();

        List<Future<Boolean>> futures = IntStream.range(10, 30000).boxed().map(s ->
                executorService.submit(() -> {
                    try {
                        valOpt.set(PERFORMANCE_PREFIX + s, "id");
                    } catch (Exception e) {
                        log.error("", e);
                    }
                    return true;
                })).collect(Collectors.toList());

        for (Future<Boolean> future : futures) {
            try {
                future.get();
            } catch (Exception e) {
                log.error("", e);
            }
        }

        return Result.success(System.currentTimeMillis() - begin);
    }

2.響應式代碼

   // 使用reactive 耗時：6.7s  使用mysql-2線程池 redis-2線程池
    @PostMapping(value = "reactiveTestWrite2Redis")
    public Mono<Result<Long>> reactiveTestWrite2Redis() {
        long start = System.currentTimeMillis();
        ReactiveValueOperations<String, String> valOpt = r2redisTemplate.opsForValue();

        final Mono<Long> longMono = Flux.range(10, 30000)
                .flatMap(s ->
                        valOpt.set(PERFORMANCE_PREFIX + s, String.valueOf(s))
                                .flatMap(e -> {
                                    //log.info("set id={}, done", s);
                                    return Mono.just(s);
                                }))
                .count()
                .flatMap(s -> {
                    log.info("finish ...");
                    return Mono.just(System.currentTimeMillis() - start);
                });

        return Result.success(longMono);
    }

性能截圖：

非響應式

響應式

總結：

在這個場景下，性能快2倍副砍。如果是io延遲更高的情況下，差距會更加明顯。