This message was deleted.

Nevermind; fixed by adding this into the k3d config;

env:
  - envVar: KUBECONFIG=/var/lib/rancher/k3s/agent/kubelet.kubeconfig # same as `--env 'bar=baz@server:0'`
    nodeFilters:
      - agent:*

Should this be done by default? Not sure what the agent is trying to fetch

Those are K3s logs, so k3d shouldn't have any stakes in this. The workaround you have there is also nothing I would consider to build into k3d by default if not necessary. @creamy-pencil-82913 do you know what's going on there?

sure, but I would expect k3d to spin up non-error-spewing k3s pods by default? or is that not realistic? 😉

Is this actually affecting any functionality of your cluster? That'd be a different story then.

well, I had pretty high k3s CPU-usage; so my first go-to was to check the logs to see if its stuck doing anything

That's a good take. Do you want to put in a PR for this?

but, I'm not a 100% sure what the actual decrease it, actually working on scripting to up identical environments, once I get everything working 100% I'll bench identical environments with/without the fix and do a PR if its significant

That would be pretty cool 👍

well, I guess I could up logging in k3s to see what its actually trying to do and failing at, as thats kinda interesting as the nodes were functioning just fine