So this is a bit more to understand how APM works,
How come Elastic APM agents are standalone agents from Elastic Agent?
I see other APM tools using their "App Server Agent" to be able to detect the applications on the server and then run the instrumentation based if the user decides to monitor that application.
This makes it easier for the user since its under one umbrella and they can easily identify what applications can be monitored on the server.
It's obviously technically feasible and does make it easier for initial setting up and testing and for very small deployments and some types of applications. But on larger systems you'd have very complex rules deciding which applications to inject agents to, and most language agents can't be injected dynamically, they need to be declared on startup which means this mode would be less useful. Ultimately there isn't that much demand for this. We have tested doing it for Java agents (which allows dynamic agent attach) and it works fine, but we're not moving forward with this feature at this time
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.