We are seeing many devices where profiles are not getting deployed properly. Because of this issue, we are seeing many devices that are collecting metrics from an 'Unknown probe.' When we check that in the database, we are not able to find that qos information. cdm profiles that were deployed seemed healthy but had no Disk metrics, for example QOS_DISK_USAGE_PERC. These unknown metrics are coming from the UIMAPI.
Environment
DX UIM 23.4 CU2
cdm 7.20
mon_confg_service (MCS)
cdm_mcs_templates 7.20.1
Cause
There were two cdm.exe instances running on the same system even after the robot was manually restarted. Restart of the robot had no affect because the other instance was stale and the controller couldn't stop it as it was not the parent of it... but one instance was closed and started when the robot started.
Resolution
In the environment for this one robot, there were two instances of cdm.exe running.
After deactivating the cdm probe, killing the extra cdm.exe instance, and then reactivating the probe, everything worked fine.