CPU threshold alerts

[Ben Butler]

I have set CPU thresholds on a single server, SNMP is configured correctly, and sending a test trap delivers the alert to the CMS. the server is on the other side of a firewall, which allows two way SNMP traffic.... tested by sending a test trap and using SNMP polling.

The problem is this;

the CPUs go well above the warning and critical thresholds but no alerts can be seen on the CMS. any ideas why its not working, other alerts are coming through?

[Rob Buxton]

CPU Threshold breaches are not triggered by spikes, there needs to be a few iterations before the threshold is considered to be breached. Not too sure of how long before it reports the condition.

[Ben Butler]

Thanks Rob, I'll run the CPU stress tool for a few different periods and see what happens, i'll let you know the results.

[Steve Kadish]

With the default SIM and Management Agent settings, you won't get an alert until the threshold has been exceeded for 30 minutes. This is determined by the sampling interval set in the Management Agents Control Panel on the monitored node (which you can control) and the sampling interval used by SIM (which I'm not sure if you can control).

- Steve

[Ben Butler]

answer recieved