MCS CDM template metrics enablement creates mismatch between OC and IM CDM GUIs

book

Article ID: 222187

calendar_today

Updated On:

Products

DX Infrastructure Management

Issue/Introduction

When deploying a CDM Enhanced template the following behaviour is seen.

When Activating CDM single CDM metric via MCS additional metrics are also activated in the probe's gui in IM. There is a mismatch between what I can see in the OC MCS template. 

Is this expected? 


Example:

1. Have cleaned CDM probe (no qos enabled) on 2 robots. 

https://api-broadcom-ca.wolkenservicedesk.com/attachment/get_attachment_content?uniqueFileId=in/N1cVw+8WEqTnh1LKdpQ==

 

2. Deploy 1 CPU Enhanced profile enabling only 1 "Total CPU Usage" QOS. 

https://api-broadcom-ca.wolkenservicedesk.com/attachment/get_attachment_content?uniqueFileId=OvXUXmxm9nCsWeENgu6ecQ==


3. Wait for the profile to deploy and see the result. 

4. All of these is activated instead of just Total CPU Usage:

 

Also,  When Deactivating the metric from MCS CDM profile the change is not reflected in the probe configuration. 

Cause

Working as designed. Data is filtered out by the spooler based on plugin_metric.cfg 

Environment

Release : 20.3

Component : UIM - MON_CONFIG_SERVICE

Resolution

Even though additional metrics are enabled at a GUI Level in the CDM configuration in IM, only the enabled metrics will be actually stored into the DB. 

When enabling a single QoS object for CPU and after the successful deployment of the monitoring profile to the target device, the probe cfg is updated to enable all of the CPU QOS objects and this can also be seen in the probe configuration GUI via IM or AC.

This is by design across all of the probes when using Enhanced MCS profiles, at monitoring probe level it has permission to send all of the group metrics for CPU however, once those metrics reach robot level the spooler uses the plugin_metric.cfg to filter out any unselected QOS from the MCS UI. The spooler probe is acting as a filter to prevent any unselected metrics from reaching the UIM database. 

Here is an overview of the workflow: 

- setup cdm enhanced profile configured 
- cdm probe is pushed out to the target device
- cdm cpu enhanced profile configured
- enhanced monitoring profile pushed out to the target device and the plugin_metric.cfg is updated with the "selected" mcs ui qos. 
- all metrics by default will be enabled at probe level for the group of metrics (e.g. CPU)
- metrics are fetched on the probe polling interval 
- robot spooler will then check the incoming metrics against the plugin_metric.cfg and filter any unwanted/unselected qos. 
*At this point the spooler is acting as a filter to ensure only the mcs ui metrics are sent to the db and visible for the device*

Attachments