We have an older agent in our environment that is v9.5 instrumented on a JBoss Fuse Fabric environment running on RHEL. We have an issue with one of our applications that results in traces being kicked off every few milliseconds which is impacting the overall JVM (we'll see the JVM disconnected off and on from the APM Infrastructure).
Is there a way to completely turn off transaction tracing within the agent.
We have set the following, is there anything we are missing as we still seem to get get traces and the JVM has been recycled after the agent changes.
At this time we cannot update the agent; looking for what we can do with the current 9.5 settings. Thank you for your anticipated help!
####################### # Transaction Tracer Configuration
# Configuration settings for Introscope Transaction Tracer i
Would fellow admins, CA staff and partners have suggestions for Razib on next steps?
If you look at the traces in trace view, what type are they ?
Trace Type: Normal or Trace Type: Sampled
That would help to determine the origin, especially if they would be majority 'sampled', sampling is an easier thing to control, I can see you have configuration for it there but would have to check if the EM can still send a request for sampling even so.
But Normal covers most other things, we don't have a trace type for automatic/deep things by itself.
Or if the traces show any AutoTrace Trigger Description, it can be something like an error, new transaction path, threshold breached (related to DA)
Is it always the same component (method/class) in the traces or a variety of methods/classes ?
or under errors tabs, if there are a lot of repetitive errors there, either see if the errors can be fixed or reduced, or set configuration in the agent profile to ignore the errors if the errors are 'expected'