Linux metrics

All network and general metrics are native without any computed metrics. The following section lists all the Linux® metrics:

Category Description

CPU

Provides you how much work your CPU is doing and how it’s being used.

For example, myMachine|CPU|cpu_user.

DiskFree

Shows how much free space is left on your disks (storage drives). Each mounted directory has a separate section.

For example, myMachine|DiskFree|myMount|df_free.

Diskstat

Provides detailed information about how each disk is being used.

Load

Provides you the information about the load statistics for a particular node. This is different from CPU usage as it shows the overall demand on the system.

For example, myMachine|Diskstat|myDisk|disk_active_ios.

Memory

Shows the information about the memory statistics for a particular node.

For example, myMachine|Memory|mem_active and myMachine|Diskstat|myDisk|disk_active_ios.

Netstat

Shows statistics about network connections (sockets, ports, and states) for a particular node.

For example, myMachine|Netstat|ns_remote_bytes_rand skstat|myDisk|disk_active_ios.

Network

Measures actual network usage of data in or out on network interfaces.

For example, myMachine|Network|myInterface|netdev_bytes_r.
TopProc sensor

Collects aggregated resource consumption for the 10 most CPU consuming processes. The sensor is different from other performance monitoring sensors. The statistics can be checked by using the mmperfmon report top command instead of with the mmperfmon query command. For more information about the mmperfmon command, see the mmperfmon command.

For example,
 mmperfmon report top -b 60 -n 1
Top values in format:
{process name | PID} {cpu per mil} {memory per mil}

Time                 node-21
2023-10-30-18:24:00  mmsysmon.py 28 13
                     pmcollector 13 8
                     tuned 12 3
                     mmfsd 12 107
                     java 9 49
                     pmsensors 7 1
                     vmtoolsd 5 1
                     kworker/1:1-ev 3 0
                     systemd 1 1
                     ksoftirqd/0 1 0
  • {process name | PID}: Is the process name or the PID.
  • {cpu per mil}: Is CPU usage in per mil (one-tenth of a percent) considering the process number of system CPU usage.
  • {memory per mil}: Is the memory usage in per mil (one-tenth of a percent) considering the total system RAM (sysInfo.totalram).

Infiniband

Provides information about the various performance metrics and error counters of InfiniBand statistics per interface for a node.

For example, myMachine|Infiniband|myInterface|port_rcv_errors