If Dynatrace does not detect you process automatically you can use a process detection rule and add the executable path so that Dynatrace can be pointed directly to the Apache process.
To do that go to settings -> processes and containers -> process group detection -> custom process group detection rule.
I think it depends on the actual situation.
If the Apache is the only process to be monitored on the host and the full stack agent does not work in your case, you can use the PaaS agent
However, I think with proper settings and mapping oneagent directories into the chroot and setting the LD_PRELOAD you should be able to use the full stack oneagent on your host in the chroot environment.
Thank you Julius for your answer.
Can you provide some more info on how to deploy the PaaS agent? I find documentation for Azure WebApps, Cloud Foundry, Openshift and Kubernetes platforms.
What steps should i follow in this case?
If not using the PaaS installer,
would the following settings solve my issue?
mkdir -p $CHROOT/var/lib/dynatrace
mount -o bind /var/lib/dynatrace/ $CHROOT/var/lib/dynatrace/
mkdir -p $CHROOT/opt/dynatrace
mount -o bind /opt/dynatrace/ $CHROOT/opt/dynatrace
mount -o bind /dev/ $CHROOT/dev
cp /etc/resolv.conf $CHROOT/etc/resolv.conf
Please advise also where to set the LD_PRELOAD variable.
The LD_PRELOAD is typically set from /etc/ld.so.preload and typically is set to /lib/x86_64-linux-gnu/liboneagentproc.so
So you need to either add /lib/x86_64-linux-gnu/liboneagentproc.so to the chroot and set the LD_PRELOAD correctly or you can try to point it to /opt/dynatrace/oneagent/agent/lib64/liboneagentproc.so (I did not try - it may not work).
The PaaS agent can be downloaded from this URL (replace tenant and api token)
But don't forget you have to manually upgrade the PaaS agent. For usage of PaaS agent, I'd recommend seeing the docs for Solaris OneAgent as it only exists as PaaS. (There are no docs about PaaS agent on Linux, since typically you install the full-stack agent).