Unsolved
This post is more than 5 years old
7 Posts
0
1194
December 13th, 2013 00:00
Sustained CPU & Memory Alerts
Hi All,
Is there any option to configure CPU & memory alerts with time based ?
For example if CPU of a device is above 90% continuously (90% is threshold) for 45 mins i need a alert in service assurance manager?
No Events found!
hubery
13 Posts
1
December 15th, 2013 16:00
You have a few options:
1. Use SAM notification module to control - set up NCI ( notification configuration item ) and configure the publish delay
( see detailed information in the SAM notification module user guide )
2. If you have watch4net, you can use the w4n alerting module instead of the Smarts IP-AM-PM polling and threshold.
If you are familiar with ASL, you can write a hook script, apparently, this is harder than the above 2 options.
Hubery
AlanG2
17 Posts
1
January 3rd, 2014 03:00
I've wrested with this problem for a few years. The fundamental problem is "overtime" alarms are not really supported in the underlying domains.
I would suggest W4N is the easist option here, this is how I am planning to do it.
In the past I have implemented it with a hook script and an escalation policy. When the event first notifies you put a tag in a UDF and build your filter view to hide events with that tag. You then configure an escalation policy with a level 0 that runs for the desired delay (e.g. 45 mins). At level 1 call an auto server tool that removes the tag so the event now appears in the display.
SuryaNarayana1
7 Posts
0
January 7th, 2014 03:00
Thanks huberyhe & AlanG for giving suggestion