Survey: Agent improvements

Announcements
,
1

Hello Community!
My name is Andrea and I work in the Checkmk Product Management team.
Your feedback shapes what we build next!
We’re looking at how to make the Checkmk agent easier to configure, deploy, and update and we’d like to hear where it currently gets in your way through the following survey:

:arrow_right: Agent improvements :arrow_left:

The survey takes about 5-10 minutes, and the results go directly to the team working on agent improvements.
We’re constantly striving to improve Checkmk.

Thanks,
Andrea

3 Likes
2

Hi,

thats something you should finally look into regarding agents and improve

https://ideas.checkmk.com/suggestions/306004/let-unix-agents-execute-sectionspluginslocal-checks-in-parallel

Also if you run into a timeout with a plugin or section, take the data you already collected and send it to the server instead of “loosing/dropping” all data you already collected

Cheers

D

3

Some nice ideas came up in this old thread Restructure the nix agents

I liked the idea that the agent would output JSON.

This would make it much easier to troubleshoot and compare, and filter agent outputs.

The current format with the <<<section>>> and <<<<piggibacked-host>>>>
still feels strange, even after years of working with them.

Some sections can appear mutliple times with the same name,
but with different sep() section attribute, it can be confusing e.g.:

$ check_mk_agent | grep '^<<<docker' | sort | uniq -c
     10 <<<docker_container_labels:sep(0):cached(1780471138,60)>>>
     10 <<<docker_container_labels:sep(124):cached(1780471138,60)>>>
     10 <<<docker_container_network:sep(0):cached(1780471138,60)>>>
     10 <<<docker_container_network:sep(124):cached(1780471138,60)>>>
     10 <<<docker_container_node_name:sep(0):cached(1780471138,60)>>>
     10 <<<docker_container_node_name:sep(124):cached(1780471138,60)>>>
     10 <<<docker_container_status:sep(0):cached(1780471138,60)>>>
     10 <<<docker_container_status:sep(124):cached(1780471138,60)>>>

Troubleshooting e.g. the agent output of the kubernetes special agent is a challange because of that, if you want to compare it with outputs from other clusters or versions.

1 Like