最近对线上业务进行重构，涉及到ES同步这一块，在重构过程中，为了ES 写入 性能考虑，大量的采取了 bulk的方式，来保证整体的一个同步速率，针对BulkProcessor 来深入一下，了解下 是如何实现，基于请求数，请求数据量大小 和 固定时间，刷新写入ES 的原理

针对ES 批量写入， 提供了3种方式，在 high-rest-client 中
分别是 bulk bulkAsync  bulkProcessor 3种方式。
本文主要针对 bulkProcessor 来进行一些讲述


BulkProcessor是一个线程安全的批量处理类,允许方便地设置 刷新 一个新的批量请求 
(基于数量的动作,根据大小,或时间),
容易控制并发批量的数量
请求允许并行执行。

 @Bean(name = "bulkProcessor") // 可以封装为一个bean，非常方便其余地方来进行 写入 操作
  public BulkProcessor bulkProcessor(){
        BiConsumer<BulkRequest, ActionListener<BulkResponse>> bulkConsumer =
                (request, bulkListener) -> Es6XServiceImpl.getClient().bulkAsync(request, RequestOptions.DEFAULT, bulkListener);
        return BulkProcessor.builder(bulkConsumer, new BulkProcessor.Listener() {
            @Override
            public void beforeBulk(long executionId, BulkRequest request) {
              		// todo do something
                int i = request.numberOfActions();
                log.error("ES 同步数量{}",i);
            }
            @Override
            public void afterBulk(long executionId, BulkRequest request, BulkResponse response) {
						// todo do something
                Iterator<BulkItemResponse> iterator = response.iterator();
                while (iterator.hasNext()){
                    System.out.println(JSON.toJSONString(iterator.next()));
                }
            }
            @Override
            public void afterBulk(long executionId, BulkRequest request, Throwable failure) {
								// todo do something
                log.error("写入ES 重新消费");
            }
        }).setBulkActions(1000) //  达到刷新的条数
                .setBulkSize(new ByteSizeValue(1, ByteSizeUnit.MB)) // 达到 刷新的大小
                .setFlushInterval(TimeValue.timeValueSeconds(5)) // 固定刷新的时间频率
                .setConcurrentRequests(1) //并发线程数
                .setBackoffPolicy(BackoffPolicy.exponentialBackoff(TimeValue.timeValueMillis(100), 3)) // 重试补偿策略
                .build(); 
    }

bulkProcessor.add(xxxRequest)

BiConsumer<BulkRequest, ActionListener<BulkResponse>> bulkConsumer =
                   (request, bulkListener) -> Es6XServiceImpl.getClient().bulkAsync(request, RequestOptions.DEFAULT, bulkListener);

    return BulkProcessor.builder(bulkConsumer, new BulkProcessor.Listener() {
            ****
            
           }).setBulkActions(1000)
                   .setBulkSize(new ByteSizeValue(1, ByteSizeUnit.MB))
                   .setFlushInterval(TimeValue.timeValueSeconds(5))
                   .setConcurrentRequests(1)
                 .setBackoffPolicy(BackoffPolicy.exponentialBackoff(TimeValue.timeValueMillis(100), 3))
                   .build();
   
       }

Listener
Builder
BulkProcessor
Flush
===== 华丽的分界线
BulkProcessor 实现了 Closeable --> 继承自  AutoCloseable （关于AutoCloseable 本文不做过多说明，具体的可以百度，或者等待后续）

/**
* 简单的构建，可以看到，就是一个client 和 listener 这个不会做刷新策略，
*/
public static Builder builder(Client client, Listener listener) {
        Objects.requireNonNull(client, "client");
        Objects.requireNonNull(listener, "listener");
        return new Builder(client::bulk, listener, client.threadPool(), () -> {});
    }
/**
* 所有功能的builder 实现方法
* ScheduledThreadPoolExecutor 用来实现 按照时间频率，来进行 刷新，如 每5s 
*
*/
    public static Builder builder(BiConsumer<BulkRequest, ActionListener<BulkResponse>> consumer, Listener listener) {
        Objects.requireNonNull(consumer, "consumer");
        Objects.requireNonNull(listener, "listener");
        final ScheduledThreadPoolExecutor scheduledThreadPoolExecutor = Scheduler.initScheduler(Settings.EMPTY); // 接口静态方式，来实现 Executor 的初始化
        return new Builder(consumer, listener,
                (delay, executor, command) -> scheduledThreadPoolExecutor.schedule(command, delay.millis(), TimeUnit.MILLISECONDS), //
                () -> Scheduler.terminate(scheduledThreadPoolExecutor, 10, TimeUnit.SECONDS));
    }
/**
* 构造函数 
* @param consumer 前文定义的consumer request response action
* @param listener listener  BulkProcessor 内置监听器
* @param scheduler elastic 定时调度 类scheduler  
* @paran onClose 关闭时候的运行
*/
    private Builder(BiConsumer<BulkRequest, ActionListener<BulkResponse>> consumer, Listener listener,
                        Scheduler scheduler, Runnable onClose) {
            this.consumer = consumer;
            this.listener = listener;
            this.scheduler = scheduler;
            this.onClose = onClose;
        }

   
   BulkProcessor(BiConsumer<BulkRequest, ActionListener<BulkResponse>> consumer, BackoffPolicy backoffPolicy, Listener listener,
                  int concurrentRequests, int bulkActions, ByteSizeValue bulkSize, @Nullable TimeValue flushInterval,
                  Scheduler scheduler, Runnable onClose) {
        this.bulkActions = bulkActions;
        this.bulkSize = bulkSize.getBytes();
        this.bulkRequest = new BulkRequest();
        this.scheduler = scheduler;
        this.bulkRequestHandler = new BulkRequestHandler(consumer, backoffPolicy, listener, scheduler, concurrentRequests); // BulkRequestHandler 批量执行 handler 操作
        // Start period flushing task after everything is setup
        this.cancellableFlushTask = startFlushTask(flushInterval, scheduler); //开始刷新任务
        this.onClose = onClose;
    }

 private Scheduler.Cancellable startFlushTask(TimeValue flushInterval, Scheduler scheduler) {
   
   // 如果 按照时间刷新 为空，则直接返回 任务为取消状态
   if (flushInterval == null) {
            return new Scheduler.Cancellable() {
                @Override
                public void cancel() {}
                @Override
                public boolean isCancelled() {
                    return true;
                }
            };
        }
        final Runnable flushRunnable = scheduler.preserveContext(new Flush());
        return scheduler.scheduleWithFixedDelay(flushRunnable, flushInterval, ThreadPool.Names.GENERIC);
    }
    private void executeIfNeeded() {
        ensureOpen();
        if (!isOverTheLimit()) {
            return;
        }
        execute();
    }
// 刷新线程
class Flush implements Runnable {
        @Override
        public void run() {
            synchronized (BulkProcessor.this) {
                if (closed) {
                    return;
                }
                if (bulkRequest.numberOfActions() == 0) {
                    return;
                }
                execute(); // 下面方法
            }
        }
    }
/**
*  刷新执行
*
*/
 // (currently) needs to be executed under a lock
    private void execute() {
        final BulkRequest bulkRequest = this.bulkRequest;
        final long executionId = executionIdGen.incrementAndGet();
				// 刷新 bulkRequest 为下一批做准备
        this.bulkRequest = new BulkRequest();
        this.bulkRequestHandler.execute(bulkRequest, executionId);
    }

   
/**
* 各种添加操作
*/
public BulkProcessor add(DocWriteRequest request, @Nullable Object payload) {
        internalAdd(request, payload);
        return this;
    }
	/**
	* 我们可以看到，在添加之后，会做一个操作
	* executeIfNeeded 如果需要，则进行执行
	*/
    private synchronized void internalAdd(DocWriteRequest request, @Nullable Object payload) {
        ensureOpen();
        bulkRequest.add(request, payload);
        executeIfNeeded();
    }
/**
* 如果超过限制，则执行刷新操作
*/
 private void executeIfNeeded() {
        ensureOpen();
        if (!isOverTheLimit()) {
            return;
        }
        execute();
    }
	/**
	* 这这儿，我们终于看到了 关于action 和 size 的判断操作，
	*
	*/
    private boolean isOverTheLimit() {
        if (bulkActions != -1 && bulkRequest.numberOfActions() >= bulkActions) {
            return true;
        }
        if (bulkSize != -1 && bulkRequest.estimatedSizeInBytes() >= bulkSize) {
            return true;
        }
        return false;
    }

针对数据大小的设置。在每次添加的时候，做判断是否 超过限制
针对 时间的频次控制，交由ScheduledThreadPoolExecutor 来去做监控

  public void execute(BulkRequest bulkRequest, long executionId) {
        Runnable toRelease = () -> {};
        boolean bulkRequestSetupSuccessful = false;
        try {
          // listener 填充 request 和执行ID
            listener.beforeBulk(executionId, bulkRequest);
          	//通过信号量来进行资源的控制 来自于我们设置的 setConcurrentRequests
            semaphore.acquire();
            toRelease = semaphore::release;
            CountDownLatch latch = new CountDownLatch(1);
          // 进行执行并按照补偿重试策略如果失败
            retry.withBackoff(consumer, bulkRequest, new ActionListener<BulkResponse>() {
						 	//结果写入  ActionListener --> BulkProcessor.Listener 的转换
              @Override
                public void onResponse(BulkResponse response) {
                    try {
                        listener.afterBulk(executionId, bulkRequest, response);
                    } finally {
                        semaphore.release();
                        latch.countDown();
                    }
                }
                @Override
                public void onFailure(Exception e) {
                    try {
                        listener.afterBulk(executionId, bulkRequest, e);
                    } finally {
                        semaphore.release();
                        latch.countDown();
                    }
                }
            }, Settings.EMPTY);
            bulkRequestSetupSuccessful = true;
            if (concurrentRequests == 0) {
                latch.await();
            }
        } catch (InterruptedException e) {
            Thread.currentThread().interrupt();
            logger.info(() -> new ParameterizedMessage("Bulk request {} has been cancelled.", executionId), e);
            listener.afterBulk(executionId, bulkRequest, e);
        } catch (Exception e) {
            logger.warn(() -> new ParameterizedMessage("Failed to execute bulk request {}.", executionId), e);
            listener.afterBulk(executionId, bulkRequest, e);
        } finally {
            if (bulkRequestSetupSuccessful == false) {  // if we fail on client.bulk() release the semaphore
                toRelease.run();
            }
        }
    }

    public void execute(BulkRequest bulkRequest) {
            this.currentBulkRequest = bulkRequest;
            consumer.accept(bulkRequest, this);
        }

 @Override
        public void onResponse(BulkResponse bulkItemResponses) {
            if (!bulkItemResponses.hasFailures()) {
                // we're done here, include all responses
                addResponses(bulkItemResponses, (r -> true));
                finishHim();
            } else {
                if (canRetry(bulkItemResponses)) {
                    addResponses(bulkItemResponses, (r -> !r.isFailed()));
                    retry(createBulkRequestForRetry(bulkItemResponses));
                } else {
                    addResponses(bulkItemResponses, (r -> true));
                    finishHim();
                }
            }
        }
/**
* 只针对失败的请求，放入到重试策略中
*/
   private void addResponses(BulkResponse response, Predicate<BulkItemResponse> filter) {
            for (BulkItemResponse bulkItemResponse : response) {
                if (filter.test(bulkItemResponse)) {
                    // Use client-side lock here to avoid visibility issues. This method may be called multiple times
                    // (based on how many retries we have to issue) and relying that the response handling code will be
                    // scheduled on the same thread is fragile.
                    synchronized (responses) {
                        responses.add(bulkItemResponse);
                    }
                }
            }
        }