Monitoring CPU usage

VCS includes a system attribute, CPUUsageMonitoring, which monitors CPU usage on a specific system and notifies the administrator when usage has been exceeded.

The default values for the CPUUsageMonitoring attribute are:

The values for ActionTimeLimit and NotifyTimeLimit represent the time in seconds. The values for ActionThreshold and NotifyThreshold represent the threshold in terms of CPU percentage utilization.

If Enabled is set to 1, HAD monitors the usage and updates CPUUsage attribute. If Enabled is set to 0 (default), HAD does not monitor the usage.

If the system's CPU usage continuously exceeds the value set in NotifyThreshold for a duration greater than the value set in NotifyTimeLimit, HAD sends notification via an SNMP trap or SMTP message.

If the CPU usage continuously exceeds the value set in NotifyThreshold for a duration greater than the value set in NotifyTimeLimit, subsequent notifications are sent after five minutes to avoid sending notifications too frequently (if the NotifyTimeLimit value is set to a value less than five minutes). In this case, notification is sent after the first interval of NotifyTimeLimit. As CPU usage continues to exceed the threshold value, notifications are sent after five minutes. If the values of NotifyThreshold or NotifyTimeLimit are set to 0, no notification is sent.

If system's CPU usage exceeds the value set in ActionThreshold continuously for a duration greater than the value set in ActionTimeLimit, the specified action is taken. If the CPU usage continuously exceeds the ActionThreshold for a duration greater than the value set in ActionTimeLimit, subsequent action is taken after five minutes to avoid taking action too frequently (if the ActionTimeLimit value is set to less than five minutes). In this case action is taken after the first interval of ActionTimeLimit. As CPU usage continues to exceed the threshold value, action is taken after five minutes. If the values of ActionThreshold or ActionTimeLimit are set to 0, no action is taken. Actions can have one of the following values:

none: No action will be taken and the message is logged in the VCS engine log.

reboot: System is rebooted.

custom: The cpuusage trigger is invoked.