We integrate our flows on a system that employs ConsumeAMQP 1.8. We then promote our flows as templates after evaluation, but the production system we promote to only allows ConsumeAMQP 1.7.1.g .
I’ve noticed there is a ConsumeAMQP in production that appears to maintain a thread to RabbitMQ. It never goes away. It can’t be stopped, and Terminate always seems to hang our processors. We try to avoid Terminate at all costs because it’s never once worked.
This is the first ConsumeAMQP we have promoted in this manner. Is there a way I can determine why this thread persists? It does not throw any warnings or errors. I have noticed that my messages in my Rabbit queue sometimes are not ACKed by this processor. Would promoting a flow that includes a ConsumeAMQP from a higher version nifi to a slightly lower version of nifi be a cause?
Our data does get consumed from the queue. Sometimes it is ACKed, but sometimes it is not.
Any suggestions how I can diagnose and fix this problem?