Some random notes on how to config and get netdata started.

Configuration from a slave node to an aggregation node and finally to the head node for display.

Slave:

  • netdata.conf
    • In global, 'memory mode = none' to reduce affect on the monitored node.
    • In web, 'mode = none' since we do not want to have the builtin web server running.
  • stream.conf
    • enable = yes
    • destination =
    • api key - uuidgen

Some notes:

  • api key is a uuidgen string and can be shared with multiple systems. This is different than the machines uuid which gets generated when netdata is started or from the cache file.
  • netdata does cache some information even in slave mode. This information is found the /var. If the binary is installed in a shared environment, please set '--localstatedir=' to a none share directory for each node. This is a value sent in to ./configure and not via the ./netdata-install.sh script.

Aggregation node:

  • netdata.conf
    • 'memory mode = none'
    • bind to = localhost commented off or define interface
    • In web, 'mode = none'
  • stream.conf
    • In stream, 'enable = yes' to stream the local node data.
    • In the next section that says, [API KEY] replace API KEY with the same uuidgen key from the slaves. netdata will use the following stanza of configs for all nodes connecting via this api key. This sets up netdata to receive data from other netdata processes.
    • Set 'enable = yes'
    • 'default proxy enable = yes' This starts setting up the forwarding feature
    • 'default proxy destination = IP:Port'
    • 'default proxy api key = uuidgen' This is a different UUIDGEN to communicate with the head node.

Head node:

  • netdata.conf
    • In global, unset bind to or define the bind to address. Or place netdata behind a web proxy.
  • stream.conf
    • [API KEY] set the API KEY to the second uuidgen above
    • Basically the defaults should work

At this point you should be able to get all the nodes listed and connected. Alerting is already setup and enabled.

TODO:

  • Find tune the alerting. I have not done anything with that.
  • Tune what is being collected. The amount and items that can be collected can be tuned down.
  • Investigate any plugins.

Scaling:

  • With an older version, 40 seemed to be the limit. Will update if I learn more.

Updates: If you have any more experience, or find errors in what I wrote, please correct and let me know.

Thanks

-- CaryWhitneyExternal - 2017-05-03

Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r1 - 2017-05-03 - CaryWhitneyExternal
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    HEPIX All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2024 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback