Docker Metric

[lanrekkeeg]

Is there a way to convert the docker metric cpu_used value to percentage in docker metric plugin? because this value doesn't make any sense until converted to percentage. Any formula or workaround will be appreciated.

[1] docker.0: [1571994772.00555745, {"id"=>"6bab19c3a0f9", "name"=>"postgresql", "cpu_used"=>172102435, "mem_used"=>5693400, "mem_limit"=>4294963200}]

[patrick-stephens]

Lua filter? Maybe a script that updates it that way?

[lanrekkeeg]

yes, we need to create this new plugin accordingly. I am willing to contribute from next month (July) onward.

[agup006]

@lanrekkeeg awesome if you can contribute, I would recommend looking at the NGINX metric plugin as a starter.

Additionally, you may be able to get away with using the Prometheus Scraper and configuring Docker to expose Prometheus Metrics: https://docs.docker.com/config/daemon/prometheus/

[agup006]

@lanrekkeeg would this still be something you would be open to contribute?

[lanrekkeeg]

@agup006 Yes, I had 2 exams left which I will be giving this month. After this, I will start working on this. I will look into your suggestion.

[ypo777]

Hi maintainers — any update on this?
I’m running into the same issue where Docker metrics’ cpu_used is a raw counter and not directly usable as a percentage.

Context: Fluent Bit v4.1.1, input name docker (metrics) → cpu_used, mem_used, mem_limit.

Example log:

Oct 14 12:34:19 homelab fluent-bit[3187491]: [6] docker.metrics: [[1760445258.995845997, {}], {"id"=>"04b00eb1d665", "name"=>"docker-registry", "log"=>"Oct 14 12:34:18 - docker-registry -nil", "cpu_used"=>3023971, "mem_used"=>14323712, "mem_limit"=>-9223372036854775808, "environment"=>"dev"}]

Is there any progress on exposing a cpu_percent metric (Prometheus-style) or guidance on the recommended conversion? I can work around it with a Lua filter (delta over interval and CPU count), but I’d prefer an official solution.

Thank you!