Resolution:
This is the only reason we know from the BW side.

First please understand that the situation of engine not taking any more messages and not confirming can occur if
for some reason the RVCMQ subscriber errors out WITHOUT confirming the messages.

The important configuration field is the "WorkerTaskCapacity" which is by default 10 in BW RVCMQ transport.

But when you run only engine which acts like a scheduler then the worker task capacity is set to ONE(1)
at the RVCMQ level.
This means even if one of the RVCM message errors out and does not confirm the message then the whole engine is struck.
You can simulate this easily by sending a bad message to RVCMQ sub.

If you run 2 engines with default Worker task capacity of 10 then the woker can accept 10 tasks/messages at a time.
So even if one messages is bad  itmight not have a bad impact on the engine. But eventually it can hang if all the 10 worker
tasks/threads hang with badly formatted messages

So the solution is to CONFIRM every message even though it errors out so that you do not face this problem.