Add alerting for CPU and disk utilization

We are using SNMP and feed the data to our alerting systems.  We are monitoring outbound/deferred queues, memory, load average (we can't find CPU - the SNMP counter doesn't seem to be updated),  disk usage (root, data, opt), and network traffic.

Thanks, this request has been noted.

Please do continue to add further details as necessary. 

We have been monitoring

ssCPURaw[Nice|System|User|Idle].  It appears you have integer overflow on these gauges.

To correctly plot the various ssCPURaw values, you need to compute the difference between current and previous, sum these differences and then divide each by the sum to get % utilization for each gauge.

It appears when the counter gets to 2^32-1 you don't correctly wrap to zero. Restarting the SNMP service doesn't fix this, but rebooting the box does. Also, Tech Support doesn't want to support this monitoring since the on-box MIBS don't include these counters.  However they are documented in document #  2008010311490354 SNMP OIDs and description that can be queried on Symantec Brightmail Gateway appliances

How long your system had been up before the counter reached its max value?

Can you not use ssCpuUser, ssCpuSystem, ssCpuIdle instead of their raw counterparts?  I guess these are not prone to overflow.

Regards,

Adnan

Case # 411-245-831 is now scheduled for the next 9.x release.  Turned out  only exhibits after 62 days of uptime (on an eight core processor).