Just tried an upgrade from 5.1.3.18 to 5.2.1.9 (latest stable), and NewRelic has stopped instrumenting my MySQL DB calls. Basically it counts it all as ‘JVM’ time, when on 5.1 it would split into JVM and MySQL call time (as well as detailed database query statistics).
I’ve left half the nodes running 5.1.3.18 and they’re still working OK.
Anyone else have this issue? Perhaps I will have to raise it with NewRelic
This is due to a change in recent versions of Lucee for how it handles threads internally. There is a second thread started that runs the actual CFML code to improve the ability to enforce request timeouts without hanging Tomcat. However, this internal change had the unexpected behavior of confusing most monitoring software as they only look at the original HTTP thread.
We’re discussing this internally right now, and the next Lucee release will contain a flag to disable this and probably will be disabled by default until we can find a better way to approach it. In the meantime, you won’t get the same statistics on your Lucee servers unless you back up a couple versions or wait until the next release.
Is there a Jira ticket that we can watch? Thanks!
https://luceeserver.atlassian.net/browse/LDEV-1357
It was fixed in the 5.2.2.27 snapshot build.