Auto clear alarm for CPU and Memory SpectroWatches

Products

• Spectrum

Issue/Introduction

In Spectrum we have the option to create watches for CPU/Memory so that we can raise alarms with different severities. As these watches act independently, there might be an issue when clearing the existing major alarm when you have a critical threshold alarm.

 

Environment

Release: Any
Component:

Resolution

This solution explains how the major watch CPU threshold alarm can be cleared when the CPU exceeds critical threshold . This avoids having 2 alarms in OneClick which is misleading. Expectation is that when the CPU exceeds critical threshold , the existing major alarm should get cleared.

Implement any of the following solution according to the situation

Scenario 1: When you get major and critical threshold alarms within a time gap.

Solution 1: In this scenario we can configure the events such that the critical CPU threshold watch alarm should clear the existing major alarm.

1. Logon to OneClick console
   
   
2. Open event configuration editor
   
   
3. Filter for Critical watch threshold events as per your watch configuration.
   
   
4. Go to clear options on the right bottom window
   
   
5. Add the Major Watch threshold alarm code in this table.
   
   
6. Once done go to file -> Save all in the Event configuration editor

Now when you have a critical CPU Watch threshold alarm , the existing CPU Major alarm will get cleared.

Scenario 2: The above solution will not work when you have Major and the critical watch threshold alarms at the same time. This is observed in the customer environment. Due to some reasons spectrum is not able to process the Major set alarm and the clear action at the same time.

Solution 2:

Assumption:

0xfff00001 -----> CPU Watch Critical threshold event
0xfff00002 ------> CPU Watch Major threshold event
0xfff00003 -----> Dummy Event 1 with any event message
0xfff00004 -------> Dummy Event 2 with any event message

Now follow these steps

1. Logon to OneClick console
   
   
2. Open event configuration editor
   
   
3. Filter for Critical watch threshold event as per your watch configuration.
   
   
4. Configure an event pair rule on this event such that when 0xfff00001 occurs and is not followed by 0xfff00003 with 5 minutes generate 0xfff00004
   
   
5. Now filter for 0xfff00004 and go to clear options tab at the bottom of the ECE -> Add the alarm code of Major Watch Threshold Alarm (ie 0xfff00002 ). So when 0xfff00004 occurs , the major alarm will be cleared.
   
   
6. Once done go to file -> Save all in the Event Configuration editor

The same can be implemented for Memory Utilization alarms.

Note: This Solution uses events 0xfff00001, 0xfff00002, 0xfff00003 and 0xfff00004 as examples. Your custom event id's may be different.