We have an application server that has three JVMs instrumented with APM java agent 10.5.2.52.
Of these three JVMs, one of them is behaving normally and has no issues.
The other two monitored JVMs for the same application on the same server are erratic, and constantly connect/disconnect from the APM cluster.
All JVMs use the same agent profile, agent installation, and tracers.
The erratic behavior of the agent is preventing us from monitoring the application.
The agent is showing very high auto trace activity and also there are some exceptions.
We suggest you to take the following steps.
1- Disable transaction trace sampling (on agent side in IntroscopeAgent. profile)
2- Disable auto trace if differential Analysis is not used.
add the following property in IntrosocpeEnterpriseManager.properties (on EM Collector that these JVMs reporting metrics) introscope.enterprisemanager.baseline.max.metrics.baselined=0