Global CPU utilization

[gguest64]

Hello community,

It appears that Zabbix does not provide a metric for Global CPU utilization (i.e., user + system + nice) as found in other monitoring systems (I am not referring to load average).

Could you please advise on the best method to obtain the overall CPU utilization ?

Thank you for your assistance.

[daBeast]

Looks like my Linux systems report in a CPU Utilization item:


This is what it appears on one of my system (graph mode)


I found this in the latest values area - but if you wanted it on a dashboard, you just have to create it.


Here is the setup - just pick your system and the CPU utilization

[gguest64]

Hello,

Thank you for replay. I add a specification, I use Zabbix 1.

The global CPU utilization with "all" seem to be a metric available only on Zabbix 2 :

• Alias=cpu.util:system.cpu.util
• This allows use the cpu.util key to get CPU utilization percentage with default parameters as well as use cpu.util[all, idle, avg15] to get specific data about CPU utilization."
• https://www.zabbix.com/documentation...2?hl=CPU%2Ccpu

On Zabbix 1, with don't have "All" :

• system.cpu.util[<cpu>,<type>,<mode>,<logical or physical>]
• type - possible values: user (default), idle, nice, system, iowait, interrupt, softirq, steal, guest (on Linux kernels 2.6.24 and above), or guest_nice (on Linux kernels 2.6.33 and above);
• https://www.zabbix.com/documentation...ystem.cpu.util

Thank you !

[cyber]

exactly same example for "all" is on the page of classic agent.. And it is about using aliases...not about parameters...Item is identical for classic agent and agent2.
"All" in parameters refers to "all cpu-s". https://www.zabbix.com/documentation...ystem.cpu.util

For you I can only recommend to gather all 3 of your needed parmeters (user, system, nice) and create additional calculated item for end result...