Does the field "system.cpu.user.pct" represent the server usage? and what are the contributing fields for the same

I have implemented ELK stack for logging solutions and aslo set up metricbeat for measuring server performance. I stumbled on system module's "system.cpu.user.pct" which is the system's cpu usage. Now, I need to understand what are the other fields that contribute in this value. I have other fields such as "system.cpu.cores","system.cpu.idle.pct","system.cpu.iowait.pct","system.cpu.irq.pct","system.cpu.nice.pct","system.cpu.softirq.pct","system.cpu.steal.pct","system.cpu.system.pct","system.cpu.user.pct" under the metric set cpu. But most of them have value "0". Although i will run a stress test to see if these 0 values change. I am trying to understand:

  • Does the field "system.cpu.user.pct" represent the server usage
  • If yes, what fields actually contribute to the changing values of "system.cpu.user.pct"

cpu.user means all processes consume the cpu in their user mode, and cpu.system means all processes consume the cpu in their kernel mode. Usually, user+system+idle should be close to 100%.

You can search /proc/stat in proc(5) for more.

@PChou Thank you for the quick response, This answers the first part of my question. Now, can you please tell me what are the other fields from metricbeat that will affect this server usage( user+system+idle ). As in, I have other fields as mentioned above. So, with increased stress do the values change respectively thereby contributing to the server usage?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.