Hi,
I am using camel-k 1.3 on a k8s cluster running 1.19 and using
prometheus to capture quarkus metrics from the integrations/pods.
Maybe I just haven't found the correct link, but why can the number of
inflight messages be negative like -50?
Now why am I asking:
The integration blocks after 5 to 10 minutes running or after somewhere
between 2000 and 3000 exchanges completed successfully.
I actually got it to working I just do not know why.
What I had before:
activemq:topic:route1->split->direct:route2->toD:netty:tcp->direct:route3->influxdb
The netty tcp component had 2 endpoints throwing an exception ("reset
peer" and "could not connect") which I caught using doTry()/doCatch().
I am not sure where that process initially actually blocked.
What I have now:
activemq:topic:route1->split->activemq:queue(InOnly):route2->toD:netty:tcp->activemq:queue(InOnly):route3->influxdb
My guess is that the split was actually waiting for all exchanges(split)
to finish. So the problem could be within either netty or the influxdb
component. Fun fact with my working solution the inflight count number
for the netty based route is now at ~-30000 after 2 days running.
Does anybody have an idea what I should look for or that the source of
problem is/was?
Thanks,
David
