We have noticed that when you go in to ump, and say go to Performance reports designer , if you scroll down the QOS selection by Host column you will see two different entries for the same device. One is the system name and the other is the FQDN for the same device. It appears that when regular system oriented probes are used the qos metrics appear under that entry with just the System name. But if you go and add say a specialty probe say like url_response, when you go into the UMP Performance reports designer and look for that same device again there is now a FQDN hostname with just the QOS for that Url_response. This is quite tedious trying to deal with multiple devices likes this and constantly explaining to others why they are seeing two devices. How can we marry the two into just one either system name or FQDN name? We don't care which as long as its just one device showing.
You will need to find the source of the USM device entries and from there determine how to assure that the devices are successfully correlated by the discovery_server when they are discovered. You should start with the following Knowledge Document in order to determine why you have duplicate devices in the first place:
KB000034726 : How do I find the source of a USM entry?
If you cannot determine why you have duplicate entries for these devices in the CM_COMPUTER_SYSTEM table, then I would recommend that you open a support case for additional assistance.
i understood your frustration.
if you want performance data searching starting with device, USM is the tool for it.
Performance report just displays [source] field of every performance data.
it is unfortunate - as you see, [source] assignment is probe specific, some probe may use shortname while other probe uses longname.
There is a special case where a change took place, for example system1.our.domain.na to system1.our.domain.usa. In such a case or anything like it Performance Reports Designer will show both names and QoS collected before the change will be associated to system1.our.domain.na and QoS collected after the change will be associated to system1.our.domain.usa.
Moving on past the special case, as Kathryn points out this is likely to be a problem with device correlation. discovery_server is responsible for device correlation.
How Does Discovery Correlation Work? - CA Knowledge
With UIM 8.51 SP device correlation has been improved. Quick check is if discovery_server is 8.52 SP has been deployed.
CA UIM 8.5.1 Service Pack Release - CA Unified Infrastructure Management - 8.5.1 - CA Technologies Documentation
What is happening in this case In the performance reports designer the device is listed as server1 and it contains all the System performance probe metric, then there’s another entry server1.our.domain.usa and it contains metrics for only the processes probe. (QOS_process_cpu, QOS_process_mem, and QOS_process_state). The process information never appears under the server1 name.
OK so with server1 & server1.ourdomain.use both showing up, it points to a device correlation problem as noted above.
Is your robot name shortname or FQDN ?
To correct this you can do the following
1) in the robot controller make sure the robot name is selected and set to the short name
2) On the controller, MISC tab make sure the checkbox is checked for send QOS as the short name instead of computer hostname
3) in CDM on the setup General tab make sure it is checked to send shortname as QOS source.
4) once this is done stop the robot service and clear the niscache folder.
5) Go into UMP and open the SLM portlet and delete both CDM metrics for the short and the FQDN.
6) start the robot service.
The metrics should now come in only with the short name and should be listed only once and the issue should be resolved.