I have configured the jvm_monitor probe on multiple robots, under different origins, with identical profiles, monitoring a local JVM application that is running on the same server as the robot.
The Resource Name is identical on each instance of jvm_monitor because the probe is monitoring the same JVM application server at different remote sites. The profile name is the name of the application being monitored as seen below:
This is configured the same way on several different robots in my environment.
However, I am unable to see the data from this probe in Operator Console/Performance Report Designer because it only ever shows me a single instance which has the application name as the "Host" value:
I can see that the instances are separated correctly in S_QOS_DATA:
How can I find and select specific instances for reporting?
This occurs because the jvm_monitor probe uses the profile/resource name as the "Source" value for QoS metrics, so each profile (even from different instances of the probe) needs a unique name.
Broadcom recommends appending the robot name, IP address, hostname, or other unique identifier to the Resource name.
Currently this is the only way to allow identical configurations across multiple instances of the probe without conflict.
Some probes (most notably the net_connect probe) can override the QoS source with something else (like robot name or IP address) to work around this issue; jvm_monitor does not currently have this capability as of February 2024 but an Idea has been posted for this here.