We have recently experienced a few incidents in which application updates/deployments have resulted in previously detected services becoming inactive and new services coming online. Unfortunately, this has resulted in us having to re-implement a large number of service-level configurations (e.g., request attribute extraction rules, request naming rules, key request tagging). As a result, I would like to document all of the things that will result in new services being detected by Dynatrace, so that application teams will be able to evaluate whether upcoming deployments will “break” the existing monitoring.
I found some useful information at the following URL, but I have not been able to find clear enough information related to process group detection and what changes will result in new process groups being detected (and hence new services being created).
I assume that any changes to the singular properties shown in the process properties would result in a new process group being detected. For example, changes to the configuration affecting attributes like EXE path, JVM version, Command Line Arguments etc. would presumably have to result in new process groups being detected (since the GUI shows only 1 value for these attributes). Would that be a valid assumption?
For example, for WebLogic I assume that changes to any of the following would result in a new process group:
Command line args
JVM vendor or version
Conversely, I assume that plural attributes like “Ports” would not change the process group composition… Is this also a valid assumption?
This is very interesting, I would recommend opening a support ticket to see what can be done to resolve this issue/concern that you are having. This is the first time that I have heard of this as an issue.
Its not a defect for which I can open a support ticket... the behaviour is expected... the only question is one of documentation... i.e., exactly what changes result in new Process Groups and, hence, services.