Stop filling monitoring queues when processor fails

[jbescos]

Related to this #4152 but there are more things to investigate.

This PR is about not produce new events in the queues when there is no consumer.

[jansupol]

Can CountdownLatch be used instead of sleep? Sleep tends to create intermittent failures.

[jbescos]

In the first Thread.sleep we can do it because it is about waiting till something happens.

But in the second we cannot do it, because nothing should happen. So we need to wait explicitly and check.

[jbescos]

I was trying to add a CountDownLatch for the first one but it is not easy. There are 2 ways:
Easy one. It is about checking the queue till the size is 0:

    private void waitTillQueueIsEmpty() throws InterruptedException {
        final CountDownLatch latch = new CountDownLatch(1);
        ScheduledExecutorService scheduler = Executors.newSingleThreadScheduledExecutor();
        scheduler.scheduleAtFixedRate(() -> {
            Response response = target("/example/queueSize").request().get();
            int queueSize = response.readEntity(Integer.class);
            if (queueSize == 0) {
                latch.countDown();
            }
        }, 0, 10, TimeUnit.MILLISECONDS);
        boolean wait = latch.await(TIMEOUT, TimeUnit.MILLISECONDS);
        scheduler.shutdown();
        assertTrue("The queue was not empty in the given time", wait);
    } 

But this doesn't work always, because it could happen the scheduler consumed all the events and the size is 0, but new events are coming.

The difficult option:
Pass the CountDownLatch reference to the Runnable in the Scheduler of MonitoringStatisticsProcessor, so it will count it down every time an event is consumed. This is tricky because the way it is done it is not possible. I would need to make some changes in MonitoringStatisticsProcessor to achieve this.

I think it is not necessary to over complicate it, because the scheduler is running every 1 millisecond and it is very unlikely that during 1 second it was not able to do the work. But let me know what you think.