distributed cpu monitoring

[dukenukem]

Somebody please advise!

I have configured on zabbix to monitor server with 8 CPUs via SNMP. Each CPU has its own item and trigger configured. In my case I had to attain an email and SMS to be sent when the load of the CPU is reaching 70 percent. Each CPU has its own SNMP OID. It's working perfectly and I get an email and SMS each time when the load of each CPU is over 70 percent. I want the CPU threshold to be calculated on the total CPU, not separately for each of the 8 CPUs and I don't know if there is an opportunity to attain this performance. Maybe somebody knows how to do it?
Thank you.

[MrKen]

You could create a 'calculated item', which adds the other 8 values together, like:

last("MyHost:MyKey[cpu1]")+last("MyHost:MyKey[cpu2]")+last(etc., etc. . .

then add a trigger to that item.

http://www.zabbix.com/documentation/1.8/manual/config/items#calculated_items

MrKen

[dukenukem]

Thank you very much MrKen. It's working great.

[dukenukem]

Continue to this question

Continue to this question, is there a possibility to create an average of all CPUs participating in this calculated item? Meaning when one CPU is working hard 100% but all the rest are on 0%, ZABBIX will perform an average for all CPUs and just in case the sum average of all CPUs will go over 70% let's say, admin will get a warning message?

Thank you very much in advance.