Could I get DISKIO per PROCESS from metricbeat dashboard for kibana?

(alan) #1

Recently I tried the metricbeat,found it very nice,it provides a sample dashboards for kibada,so we can easily build up a dashboard for SYSTEM,but there is a pity that i can't find a table about DISKIO for per PROCESS,for example,I want to find which process cause a heavy DISKIO,but i can not get it from the dashbord.

We can see from pic about ,dashbord have Top Processes By CPU and Top Processes By Memory,but no Top Processes By IO.
Cound someone tell me how to get the information abort IO for per PROCESS?

I use 6.2.4 metricbeat and 6.2.4 ES and kibana



(Carlos Pérez Aradros) #2

Hi @alantany,

You can retrieve IO per process from system.process.cgroup.blkio... metrics, check to see how to enable and the expected output.

Best regards

(alan) #3

I had noticed the system.process.cgroup.blkio metric,but there have no information collected for it,there is no value for it in the ES .

I think it's set to true by default,I didn't changed it.

here is my metircbeat.yml:

reload.enabled: true
reload.period: 10s


  • module: system
    • cpu # CPU usage
    • filesystem # File system usage for each mountpoint
    • fsstat # File system summary metrics
    • memory # Memory usage
    • load # CPU load averages
    • network # Network IO
    • process # Per process metrics
    • process_summary # Process summary
      #- core # Per CPU core usage
    • diskio # Disk IO
    • socket # Sockets and connection info (linux only)
      enabled: true
      period: 20s
      processes: ['.*']

process.cgroups.enabled: true

host: ""

enabled: true
hosts: [""]
#index: "metricbeat-%{[beat.version]}-%{+yyyy.MM.dd}"

my server is
Linux version 3.8.13-68.3.4.el6uek.x86_64 (mockbuild@x86-ol6-builder-06) (gcc version 4.4.7 20120313 (Red Hat 4.4.7-3) (GCC) ) #2 SMP Tue Jul 14 15:03:36 PDT 2015

Thank you very much!

(Carlos Pérez Aradros) #4

Hi again,

Sorry, I've been researching this more, and it seems we only support these metrics for processes running on cgroups (you could potentially use cgexec to run them, but that's not what you are looking for).

I saw you opened a new issue in Github, could you please add more details there? This way we can track interest and allow you (and others) to suscribe for updates.

Best regards

(Carlos Pérez Aradros) #5

Also, I see there are some previous efforts here:

You can link that on your issue, for awareness

(system) #6

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