We have a base springboot docker image with the New relic agent. This image is then extended by other teams to build specific Java services/applications. These services are running in docker on EC2 instances. NR is unable to recognize HTTP traffic for some of these services (the HTTP calls are invoked via a jar file). If I add custom instrumentation for these specific classes the agent seems to pick these up (still doesn’t provide the level of detail it is for the other HTTP transactions but a start). We have high security turned on in our account, so don’t think I can use the custom instrumentation ui or do it via the web UI. Before we go down the path of adding custom instrumentation had a few questions –
- Do we have to add all the method names in the custom-instrumentation.xml file or can I add the class name and get NR to recognize the different methods in it?
- Since we are leveraging a base image for the new relic installation, if we end up adding the custom instrumentation in the base image, it will flow down to applications which do not have the classes specified in the custom instrumentation file. Will that cause errors? Will that cause performance impacts to any of these services which don’t really leverage these classes?