Anyone know the elastic_apm_application_packages value for Jenkins?

I have no idea where/how to find this value, I've exhausted my Google Fu capabilities.


Hi @trudesea

Did you simply try not setting it... it is actually not required...its a good practice but not required.

Also, I just downloaded the jar and took a look

You also might think about some of the automatic profiler

And I assume you looked at this?

Thanks, not a Java guy so I was a bit lost there.

We implemented OpenTelemetry, but there is a bug in Elastic (confirmed by support ticket) where the Java/CPU/Heap, etc metrics are not populated in APM. The data is there, but APM doesn't see it. I would prefer to use Opentelemetry, but exploring all my options and having APM as an init container in ECK was an option provided by the support rep.

What version of the stack?

Try our APM Agent just leave the packages out for a first try.

Yeah, if I recall they're not properly mapped to our fields for the OOTB experience but you can build a dashboard in 15 to 30 minutes. You just have to use the open telemetry fields.

We are on 8.5.2.

Yes we built a dashboard for that. The Jenkins plugin has some nice QoL capabilities and I'd like to keep using it. The other reason I'm trying the agent is that our Jenkins master shortly after configuring the plugin, started restarting, but we are also in heavy Jenkins pipeline development at the moment We removed the plugin and are monitoring it to see if that behavior stops. I'm trying the apm agent as a backup if it is indeed the plugin causing issues.

So did you try the elastic agent without setting the packages it should work... then you can start narrowing it down

I'm working on that now

Seems to be working, more testing to be done...ty

1 Like

Cool... Let us know if / what you did to get it working ... or was it just magic :slight_smile: and/or share any tips for the next person.

I just put a random value in for the elastic_apm_application_packages environment variable. I did not attempt to run apm without knowing what I needed to set it to. This document says it's part of the minimal configuration and important. I didn't want to waste my time if I didn't know the proper value.

Since you said it isn't required, I tried just setting it to something and everything works.


Ok so after more testing, we aren't getting expected trace names or logs, we are getting 2 traces:

That's it and traces go no deeper than that. With OT we were getting the job names, logs, etc

We do have other elastic apm agents running on JIRA/Liferay and they are working correctly.
Only thing I can think of is something having to do with running it as an init container in the Jenkins master deployment, but it obviously is seeing somethings such as java heap, cpu etc.


Another alternative here would be to use OTel agent directly with APM Server, APM Server has an OTLP endpoint, so evey OTel library/agent can use it as destination.

The complete documentation is available here, the suggestion above is available under the OpenTelemetry native support header.

This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.