We are facing a problem where no cdm QoS metrics are being collected across our robots. Looking into the issue, we found that the plugin_metric.cfg files do not contain the CPU, Memory, and Disk profiles anymore. These profiles derive from group profiles, but the file now only contains "$setup_profile$ = 1" entries for each profile ID.
The Monitoring Configuration Service (MCS) might be failing to properly deploy or synchronize the full metric configuration to the local robot cache, leaving the configuration files in an incomplete state that prevents data collection.
The following steps outline the diagnostic and resolution process to restore QoS collection:
Perform nis_cache Reset To verify if local cache corruption is the cause on a specific robot:
nis_cache_clean callback. _reset_device_id_and_restart callback.
Full Probe Reinstallation: If redeploying profiles does not fix the inconsistency, a manual removal and fresh deployment of the cdm probe is required: