Existing CPU Threshold Alert Not Triggering Despite High Utilization in DX NetOps

Products

• Network Observability
• CA Performance Management

Issue/Introduction

Users may encounter a situation where a configured threshold profile in DX NetOps Performance Management fails to trigger an alarm for a specific device, even when the following conditions are met:

• The CPU utilization graph consistently shows values above the defined threshold (e.g., >99%).
• Creating a new/duplicate threshold profile with the same rules and group assignment triggers the alarm successfully.
• The metric is confirmed to be polling correctly.

Environment

Performance Management  - all supported versions

Cause

This issue is typically caused by an existing event in the database. If an event is already active for a specific device and rule combination, it will not re-assert or trigger a new event until the previous state is cleared. 

Resolution

Step 1: Verify Active Event

Check the DX NetOps Console for any existing active event for the device. If an event is already open, the system will not generate a duplicate notification. 

Step 2: Clear the Threshold State (Group Toggle)

If no active event is visible but the profile still fails to trigger, perform a group toggle to force a reset:

1. Identify the Group that the non-working threshold profile is associated with.
2. Temporarily remove the affected device from that group.
3. Wait for the system to process the removal. While removal is often immediate, waiting until after the daily 3:00 AM maintenance cycle ensures all database states are purged.
4. Re-add the device to the original group.

Step 3: Verify the Alert

Once re-added, monitor the device. The threshold engine should perform a fresh evaluation of the current metrics and trigger the alarm as expected.