Our organization is running a .net application using rebus esb for executing business workflows. The system is processing a lot of messages with a number of handlers on a single queue.
The number of workers is set to 20 and concurrency is set to 20 while prefetch count is 10. At the start the system runs very well but later after some minutes the number of unacknowledged messages starts to accumulate and they just keep increasing.
If I restart the service then I find that the messages get processed but soon later they start accumulating. I have attached a screenshot from RabbitMQ Server that shows unacknowledged messages in the endpoint queue.
In the screenshot above you will notice that there are 5,902 messages that are not acknowledged.
Is there any way to change this behavior in Rebus?