Collect ideas for additional diagnostics checks

[jandubois]

Initial list copied from the ideas in #2319

• Are the bundled utilities accessible from a user's shell?
• Are the files under ~/.docker/cli-plugins symlinks to ~/.rd/bin?
• Has the docker context been set to rancher-desktop (or not present, if the symlink in /var/run/docker.sock is in place)?
• Is the current kubectl context set to rancher-desktop?
• Is the API endpoint responding properly (docker socket on the host, containerd socket inside the VM, k3s socket on the host)
• (macOS) Is the external cluster IP on the bridged network?
• Does the host have an active internet connection (to pull k3s versions, and container images)?
• Is the docker-credential-* helper configured and working? Make sure it is not set to none.
• Are the log files growing beyond a reasonable size?
• Are we almost out of space inside the data volume?

[jandubois]

Feedback on Slack:

Kubernetes is using context kind-argo instead of rancher-desktop.
What does this mean?

From a usability perspective, it would be great to either have that hint in the warning itself or perhaps a link to a FAQ or documentation on the rancher desktop site.

I do have multiple kube contexts available, so now that I know that that’s what it’s talking about, it makes sense.

So yeah, we should include more info, e.g. a reference to ~/.kube/config and eventually a "fix action" to change the context to rancher-desktop.