Missing Metrics Button in inCluster Deployment

[SquupS]

Hi,

I was really looking forward to the funtionality, displaying Prometheus Metrics in Headlamp. I have now updated to the latest v0.21.0 in my cluster but I don't have the button mentioned on you docs. Speaking of that one

Am I missing some config or is this only available in DesktopInstallation.

Thanks and again, great work

[joaquimrocha]

@SquupS , you're right. It's only available in the desktop version, but this plugin should also be in the official image. So we'll look into it and try to have it available in the web version for the next release.

[joaquimrocha]

BTW, adding plugins as a separate image should also not be complicated.
We have a guide on how to do that here: https://headlamp.dev/blog/2022/10/get-up-to-speed-deploying-headlamp-with-plugins/

[joaquimrocha]

This is now available in 0.22.0.

[mabrarov]

Hi @joaquimrocha,

Just FYI: metrics (at least CPU and memory consumption) of workloads is available in Kubernetes Dashboard with side container and when Kubernetes Metrics Server is deployed.

It looks like Headlamp also works with Metrics API (implemented by Metrics Server), but only for CPU and memory resources of compute nodes. I'm not sure if there is any existing / opened GitHub issue (feature request) for Headlamp to use Metrics API to show resources consumed by workloads (there are some issues which seem to be about the same, like #1380, #173, #1002 and #1).

My team considers Headlamp as replacement for Kubernetes Dashboard. So far everything looks good (and if something is missing then Headlamp team quickly adds / fixes that) - only some minor (rarely used) features are missing and lack of metrics of workloads is one of them (nice to see, but not required for real work).

Thank you.