search cancel

OC Metrics does not display any metrics in the Select Metrics list at Group or Device level

book

Article ID: 213213

calendar_today

Updated On:

Products

DX Unified Infrastructure Management (Nimsoft / UIM) Unified Infrastructure Management for Mainframe

Issue/Introduction

When creating a new metric view, OC (Operator Console) does not display any metrics in the "Select Metrics" list at Group or Device level.

Metrics are being collected, and the standard "golden metrics" like cpu/memory are displayed on the main device page, but no metrics whatsoever are showing up in the list to create a view.

Environment

Release : 20.3

Component : UIM OPERATOR CONSOLE - METRIC VIEWER

Resolution

This is often caused by some issue with the table S_QOS_SNAPSHOT.

The query to populate metrics only shows those metrics which have been collected within the past 24 hours (as this is the default for a new metric view) and this is determined by querying S_QOS_SNAPSHOT which contains the most recent samplevalues for all QoS in the database.

There are a few things to check here:

 

1. Validate if the suspect UIM instance only has a few robots and small number of metrics being collected.

2. Run the query below to confirm if the data_engine is populating current QOS data into the S_QOS_SNAPSHOT table

select max(sampletime) from S_QOS_SNAPSHOT

In this scenario the output showed a time stamp from several days back

3. Check the hub_bulk_size key on the data_engine .cfg and confirm if value is set to a high value, example 2000.

4. Revert the hub_bulk_size key value to a lower value,  20 or 50

Wait several poll cycles and validate if the Select query in Step 2 now returns a current day/time

Check Metrics in OC and now metrics being collected should be displayed

Note: The hub_bulk_size key should be set to 2000 for UIM Instances that meet the sizes in the below link:

https://techdocs.broadcom.com/us/en/ca-enterprise-software/it-operations-management/unified-infrastructure-management/20-3/installing/pre-installation-planning/ca-uim-sizing-recommendations.html

 

If this is not the issue, you should set the loglevel on data_engine to level 3 or higher and also add/change the following keys in data_engine.cfg:

log_lsv_rows = 1
lsv_sleep = 3

In most cases lsv_sleep is already present and set to 3, but if it is missing, you should add it.   log_lsv_rows is likely present and set to 0 already, so you need to flip this to "1" to enable logging of the process that updates S_QOS_SNAPSHOT.

Watch the data_engine log and every 3 seconds you should see an attempt to update the table.   If there is any issue with the database or the table itself it will be revealed here.