Additional CPU – Kafka

The Additional CPU metric group contains per-node metrics regarding CPU. The metrics available are:

• iowait
• system
• user
• irq
• softirq
• idle
• nice
• steal
• guest
• guest_nice

Iowait

CPU time the I/O thread spent waiting for a socket ready for reads or writes as a percent. High iowait times could indicate high disk utilisation.

You can also get this metric through the Monitoring API with the query param metrics=n::cpuiowaitpercent

System

Percentage of processes executing in kernel mode

You can also get this metric through the Monitoring API with the query param metrics=n::cpusystempercent

User

Processes executing in user mode, including application processes. User metric can be an indicator of high read/write to network or large amounts of non-active processes.

You can also get this metric through the Monitoring API with the query param metrics=n::cpuuserpercent

Irq

Number of hardware interrupts the kernel is servicing

You can also get this metric through the Monitoring API with the query param metrics=n::cpuirqpercent

Softirq

Number of software interrupts the kernel is servicing

You can also get this metric through the Monitoring API with the query param metrics=n::cpusoftirqpercent

Idle

Percentage of time when one or more kernel threads are executing with the run queue empty and/or no I/O operations are currently cycling.
Extended periods of idle can indicate a not all of a stream tasks partition buffers contain records. There could also be larger times for request or network processor threads not active.

You can also get this metric through the Monitoring API with the query param metrics=n::cpuidlepercent

Nice

Percentage of processes executing in user mode which have a positive nice value. Nice values are between -20 and 19 inclusive with a default of 0. Lower values translate to a higher priority in the run queue.

You can also get this metric through the Monitoring API with the query param metrics=n::cpunicepercent

Steal

Percentage of time the hypervisor allocated to other tasks external to the one run on the current virtual CPU. It can be an indication of processes that are using processor time in larger than expected amounts.

You can also get this metric through the Monitoring API with the query param metrics=n::cpustealpercent

Guest

Time spent running a virtual CPU for guest OS’ under control of kernel

You can also get this metric through the Monitoring API with the query param metrics=n::cpuguestpercent

Guest_nice

Niced processes executing in user mode in virtual OS

You can also get this metric through the Monitoring API with the query param metrics=n::cpuguestnicepercent