Thanks Julio.
That’s right, there are 3 way to do it.
But what happens is this:
I installed oneagent using the command line: bin/sh Dynatrace-OneAgent-Linux.sh --set-infra-only=true --set-auto-injection-enabled=false
and now i´m trying to deepen monitoring. Using UI or API ways i see same result:
Activation of deep monitoring was unsuccessful OneAgent auto-injection was manually disabled on this host
I think if I enable deep-monitoring via commandline, the only way to changeit is using the command line: oneagentctl
The auto-injection is your issue here. The local setting on OneAgent using installation or oneagentctl has precedence over the setting that the OneAgent receives from Dynatrace server. It also shows you that in the UI:
You should not disable auto-injection unless it’s really required for some reason. More on that in the docs.
So in your situation, you need to apply the oneagentctl command on the hosts to enable process injection. Switching to full-stack monitoring (which indeed requires injection) can be then achieved from the UI or API.
Can someone confirm that when the setting is changed on the host to false, this is not reflected in the UI? E.g. the switch is still ‘on’ and not greyed out.