node_exporter no good network usage stats?

[Sebas tech]

(Running default node_exporter in Docker, currently with "--cap-add=ALL" and /, /proc, /sys from the host forwarded.)

There are some stats for eth0 which look useful, however the system has enpXsY as main interface, and other interfaces. No good stats for those.

Looking simply for bytes received and transferred for selected interfaces over time, so to see it just as cpu usage, mem usage, etc.

Anyone with a suggestion?

[Ben Kochie]

You need to use the Docker settings from the README

https://github.com/prometheus/node_exporter#docker

--
You received this message because you are subscribed to the Google Groups "Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/42322d21-4f06-407f-87dd-7eaeaf71347dn%40googlegroups.com.

[Sebas tech]

Tried. Same outcome. With that Docker command the enpXsY interface and others do not show up at all. Only eth0, which is not the main adapter and also the old naming.

https://rakeshjain-devops.medium.com/network-interface-naming-3183bcd27389

[Sebas tech]

Edit:
My bad, apparently adding an ip to the container revokes the net host mode or so. It captures all interface now.