Closed codefromthecrypt closed 6 months ago
also, for some reason, in the other PR only the JDK11 run of the jms-jakarta invoker test was flaking, 21 hasn't so far.
also, for some reason, in the other PR only the JDK11 run of the jms-jakarta invoker test was flaking, 21 hasn't so far.
This is very interesting ... shouldn't be JDK dependent ...
here are some notes in case helps. If you don't have time, I can swing the axe again tomorrow.
@codefromthecrypt happy to look into that (it may took me a few days), we have fought a few flaky tests in Apache CXF related to Artemis migration
@reta agree to merge this. I'm wondering if CI was having a bad day. In any case, we can revisit later as isn't currently a blocker.
opened an issue as the flake returned https://github.com/openzipkin/brave/actions/runs/7438289752/job/20236893811?pr=1395
This extracts code I started to refactor while trying to get to the bottom of this error on an unrelated PR.
@reta this is on a shared branch, if you can see anything we can do to avoid flakes in activemq, please give a try. Note the failure was in the jarkarta invoker tests, and possibly resource related in CI. If there's a way to make things leaner maybe it will pass. Also, if there'a another MQ we don't need to stick with artemis either. In zipkin and reporter, I moved to docker for reasons like this. However, since this is instrumenting, I'm not sure if we can or not.. I guess we could as we arent' instrumenting the broker... Anyway, if you have some time to give a try, please do as this flake is locking up the project. If we can use docker, yank the ActiveMQExtension from zipkin or reporter and swap it out!