Jeroen Vandevelde

Regarding the requirements 1. If all threads have processed previous messages, don't request more until you hit your time period, e.g. 10 seconds _Yes, indeed that is the main requirement._...

You are correct, that with the trick of the CountDownLatch we don't need the stop function anymore. But the pulling out the logic out of the BatchingMessageRetriever would still be...

@loicrouchon If i read this correctly and i would need to build #481 with for example process 5 messages every second. I would need to implement the BackPressureLimiter, with a...

@loicrouchon I missed the fact than that if it reads a message it doesn't subtract it from the limit. In my case where i would like to have a max...

> I think this information is only fully available in a very limited number of use cases. Most of the queue consumer implementations I had to implement in the past...

@tomazfernandes I agree with your reasoning, only two points of feedback. 1. Your suggestion and the current implementation differ from the fact that currently the `DownstreamQueueBPH` uses a percentage of...

Hi @tomazfernandes > One question - how is the `TimedSemaphore` different from Guava's `RateLimiter` and why did you choose this one? I picked the `TimedSemaphore` to stick as close as...

> Hey @jeroenvandevelde, I'm thinking we should wait for #1251 to be merged before moving ahead with this one since we might end up introducing new elements you could use...

> Hey @jeroenvandevelde, #1251 has been merged and now we can support a dedicated RateLimiterBackPressureHandler in conjunction with other BackPressure implementations by using a `CompositeBackPressureHandler`. > > This should allow...