On Linux you can monitor 2 CPU metrics.
- CPU load - this is the average load across your entire system, expressed as a ratio. See this guide for an explanation.
- CPU specific metrics - this drills down into exactly what is happening on the CPU, within each CPU and as an average.
CPU specific metrics
These are broken down to a number of different metrics under the title "CPU stats", all derived from the mpstat command.
- Guest - the percentage of time spent by the CPU or CPUs to run a virtual processor.
- Idle - the percentage of time that the CPU or CPUs were idle and the system did not have an outstanding disk I/O request.
- i/o wait - the percentage of time that the CPU or CPUs were idle during which the system had an outstanding disk I/O request.
- Interrupts - the percentage of time spent by the CPU or CPUs to service hardware interrupts.
- Soft IRQs - the percentage of time spent by the CPU or CPUs to service software interrupts.
- Steal - the percentage of time spent in involuntary wait by the virtual CPU or CPUs while the hypervisor was servicing another virtual processor.
- System level - the percentage of CPU utilization that occurred while executing at the system level (kernel). Note that this does not include the time spent servicing interrupts or soft IRQs.
- User level - the percentage of CPU utilization that occurred while executing at the user level.
Individual CPUs and average across all
Server Density will show these stats across each CPU individually. If you have multiple CPUs then this can become a very large dataset so if you are running at least version 1.13.1 of our Linux agent, you will also see an option called "Average across all".
Configuring alerting
You can be alerted when CPU usage on any of these stats hits certain thresholds. This can be done on a specific CPU e.g. CPU3 or if any of your CPUs match the thresholds. If you are running at least version 1.13.1 of our Linux agent, you will also see an option called "Average across all".
Comments