Creating Alarms based on CPU Usage Thresholds on a Model

Products

• Network Observability
• Spectrum

Issue/Introduction

Requirement to have different levels of alarm based on CPU Utilization on a Model

Example:

Critical Alarm - CPU Utilization Greater Than or Equal to 90%
Major Alarm - CPU Utilization  Greater Than or Equal to 80%, Less Than 90%
Minor Alarm - CPU Utilization Greater Than or Equal to 75%, Less Than 80%

Resolution

To achieve this, we will have to create a total of 4 watches.

I have outlined the Truewatch and Critical alarms below. Also included an output of the other alarms from Oneclick. 

1. True Watch - Used for Boolean expression that will be used for the other watches:

Expression Tab



Properties Tab

2. Critical CPU - Triggers when equal to or greater than 90%

Expression Tab: 

( NRM_DeviceCPUUtilization.# >= 90 )



Properties Tab:

Threshold Tab:

Click Browse To Set Custom Cause Code - Create New With Alarm Title

Example:

3. Can be repeated for Major and Minor CPU Alarms. I have included the expressions below as examples

4.Once watches are created mark as active.

Note **Leave existing thresholds on model enabled, we will need to disable alarm for this event code: 0x00010f09**

Tools > Utilities > Event Configuration > Search for 0x00010f09 and change Severity to None

Configured Watch Report Export:

Model typeRtr_Cisco - 4watches

Name	TrueWatch
Developer ID	0xffff0000
Author	spectrum
Last Modified Time:	Nov 21, 2024 5:43:19 PM UTC
Model Type	Rtr_Cisco
Data Type	Boolean
Expression	1
Instance	None
Active By Default	Yes
Evaluate	On Demand
Inheritable	No
Threshold	None

 

Name	MajorCPU
Developer ID	0xffff0000
Author	spectrum
Last Modified Time:	Nov 21, 2024 5:50:36 PM UTC
Model Type	Rtr_Cisco
Data Type	Boolean
Expression	( ( NRM_DeviceCPUUtilization.# >= 80 ) & ( NRM_DeviceCPUUtilization.# < 90 ) )
Instance	All
Active By Default	No
Evaluate	By Polling every 0 Days + 00:00:30
Inheritable	No
Threshold	Threshold violated if value == TrueWatch . Threshold reset if value != TrueWatch . Generate Major alarm with cause code 0xfff00001 . Alarm is com.aprisma.spectrum.app.topo.common.watch.SwUIDescriptor.n user clearable. Watch will not be reset upon user clearing of alarm.

Name	CriticalCPU
Developer ID	0xffff0000
Author	spectrum
Last Modified Time:	Nov 21, 2024 5:52:03 PM UTC
Model Type	Rtr_Cisco
Data Type	Boolean
Expression	( NRM_DeviceCPUUtilization.# >= 90 )
Instance	All
Active By Default	No
Evaluate	By Polling every 0 Days + 00:00:30
Inheritable	No
Threshold	Threshold violated if value == TrueWatch . Threshold reset if value != TrueWatch . Generate Critical alarm with cause code 0xfff00002 . Alarm is com.aprisma.spectrum.app.topo.common.watch.SwUIDescriptor.n user clearable. Watch will not be reset upon user clearing of alarm.

Name	MinorCPU
Developer ID	0xffff0000
Author	spectrum
Last Modified Time:	Nov 21, 2024 5:48:52 PM UTC
Model Type	Rtr_Cisco
Data Type	Boolean
Expression	( ( NRM_DeviceCPUUtilization.# >= 75 ) & ( NRM_DeviceCPUUtilization.# < 80 ) )
Instance	All
Active By Default	No
Evaluate	By Polling every 0 Days + 00:00:30
Inheritable	No
Threshold	Threshold violated if value == TrueWatch . Threshold reset if value != TrueWatch . Generate Minor alarm with cause code 0xfff00000 . Alarm is com.aprisma.spectrum.app.topo.common.watch.SwUIDescriptor.n user clearable. Watch will not be reset upon user clearing of alarm.