I believe it does check to make sure that certain other things like agents aren't configured already when our AppMon agent starts up and the CA one would be one of those if I'm remembering an old case I saw a while ago.
I'm not familiar with the inner workings of the agent but from what I do understand of how our agent (and I assume other agents work to some extent) work is they are doing byte code modification to the application itself to let them collect all of the data and timings which I know for our agent is safe and understood but if there were somehow two agents trying to make changes to the same code it could get messy and possibly break the application. Even with our AppMon agent and the Dynatrace OneAgent only one of those can be fully injected into a single application at any given time.
Hopefully that helps and someone can add to or correct this if they have some more information,
AppMon agents and Wily agents can co-exist in the same application but I would recommend you not do that. While we didn't see any drastic issues at my client site, we did see a lot of high overhead primarily being caused by the Wily agents. If you need some verification on this, you can always install the AppMon agent and then look at the API Breakdown (or take a thread dump) to see how much overhead is being added by Wily.
Let me know if you have any questions. We went through something similar.
Satit, All good responses, a few things to add:
Yes AppMon and Wily agent can co-exist (most of the time), but it's NOT supported. You will see a warning in our agent log regarding the presence of Wily. Very infrequently there's a conflict so I would be sure your customer knows this is an unsupported configuration, suitable only for short term overlap testing and if something breaks, there's not a misunderstanding of expectations. I would strongly advise to set an expectation that this is not suitable as a long term configuration. I've done this several times, and as Ari mentioned, use the API Breakdown to show the cost of the Wily agent. It's a powerful message that the Wily agent is an ancient technology that's just not suited to a 'low overhead' approach to APM. Typical overhead I've seen with the Introscope agent is 25-30%.
Let us know what overhead you measure. Be sure to put some load on that system to get accurate real-life conditions.