If these are consumers on a queue, are they consuming the queue's messages such that your real consumers don't get all of them? Or are all of the messages still on the queue despite these consumers having received a million messages (per your screenshot in your original message)?
How often do these consumers appear, and how long do they last? And by "load", I assume you mean high CPU? Or do you mean some other measure like disk or network? What settings are you using for message expiration? Are you using scheduled messages? Do you have any embedded Camel routes? Are there any meaningful advisory messages produced during or immediately after the period of load? I don't have any good ideas for what these consumers are, so I'm guessing at things that might possibly make use of a short-lived consumer in the hopes that one of them turns out to be related to what you're seeing. But I'm grasping at straws, so this might end with "I have no idea, but maybe try upgrading to the latest 5.15.x version and see if it's still a problem there." The only other thing I notice is that the unexpected consumer uses a prefetch buffer size of 500 rather than the standard 1000. But I'm not sure why that value would be used instead of the default. Tim On Tue, May 5, 2020, 8:57 AM bstrange <bstra...@smc3.com> wrote: > The high load could be caused by the phantom consumers and not the other > way > around. I've only been able to duplicate the issue once in testing. > > > > -- > Sent from: > http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html >
