Metricbeat filesystem usage setting in units

I have 2 ELK setup with metricbeat configured to find the filesystem usage. The metricbeat.yml is like below in both the environment. But in env-01 i could see the units (GB, MB and % for the values) and in env-02 it is in decimals. Where should I update to get the units in env-02.

metricbeat.modules:

  • module: system
    metricsets:
    • cpu
    • load
    • core
    • diskio
    • filesystem
    • fsstat
    • memory
    • network
    • process
      enabled: true
      period: 10s
      processes: ['.*']
      output.elasticsearch:
      hosts: ["hostname.domain:9200"]
      logging.level: warning

From Kibana discover page, i could see the below. Is there any thing i am missing in env-02.
Env-01

  • >  # system.filesystem.free                   199.999GB
    
  • >  # system.filesystem.free_files                   7,402,970
    
  • >  t system.filesystem.mount_point                   /var/opt
    
  • >  # system.filesystem.total                   199.102GB
    
  • >  # system.filesystem.used.bytes                   1.374GB
    
  • >  # system.filesystem.used.pct                   1.24%
    
  • >  t type                   metricsets
    

Env-02

  • >  # system.filesystem.free                   196,999,999
    
  • >  # system.filesystem.free_files                   63,956
    
  • >  t system.filesystem.mount_point                   /var/opt
    
  • >  # system.filesystem.total                   249,699,759
    
  • >  # system.filesystem.used.bytes                   62,699,040
    
  • >  # system.filesystem.used.pct                   0.251
    
  • >  t type                   metricsets
    

Are you using the same version of the whole Elastic Stack? Did you run the setup process on both instances? Looks like index patterns were not correctly installed

Yes, having the same ELK versions. Installed versions are,
elasticsearch-5.0.1-1.noarch
logstash-5.0.1-1.noarch
kibana-5.0.1-1.x86_64
metricbeat-5.2.2-1.x86_64
What you mean running the setup process on both the instances? I have completely 2 isolated ELK setup for testing and Prod environment. How to test the index patterns are installed correctly and anyway to fix it if it is corrupted? please suggest.

Perhaps you can compare existing index patterns in both setups, go to Settings > Indices and check for metricbeat index template, both should have a similar number of registered fields. These are the docs for index patterns: https://www.elastic.co/guide/en/kibana/5.0/index-patterns.html

Yes, I did verified the values in the index and it is same (in decimals). So the data is available in decimals in the elasticsearch indexes in both the environment. But when it is displayed in Kibana it is converted to percentage. I have not used any scripted field but it is showing in percentage form.

The setup process involves running the import_dashboards tool which installs the index pattern for metricbeat to Kibana. This index pattern includes information about how to format certain fields (like bytes and percentages). See Importing Existing Beat Dashboards.

Thanks Andrewkroh for the input. I will go through the links and see what it is and will come back.

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