Skip to content

Conversation

@Gacko
Copy link
Contributor

@Gacko Gacko commented Nov 1, 2025

When deploying the chart to clusters with network access restricted by network policies, you might not be able access the API server.

This PR adds network policies for the controller and kubelet plugin, so they can access the API server in such environments.

Signed-off-by: Marco Ebert <[email protected]>
@copy-pr-bot
Copy link

copy-pr-bot bot commented Nov 1, 2025

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

Copy link
Collaborator

@jgehrcke jgehrcke left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you for this patch @Gacko. At the surface, this looks good to me and I trust that it has solved the problem you were seeing.

@klueska do you also want to have a look?

@Gacko
Copy link
Contributor Author

Gacko commented Nov 3, 2025

@jgehrcke: Yes, it did.

If you're interested I can come up with Helm unit testing in a separate PR. We're using the Helm Unit Tests plugin at Ingress NGINX for some basic testing.

Additionally you could use the Helm Chart Testing tool in conjunction with KIND to deploy and test the chart.

Shall I first create an issue to have a discussion?

@Gacko
Copy link
Contributor Author

Gacko commented Nov 3, 2025

Ah, and one more thought on this PR: I chose ports 443 and 6443 because they are common, but I am not sure if it would actually be a better idea to have them configurable, since this network policy hard-codes egress to any destination via 443 and 6443.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Backlog

Development

Successfully merging this pull request may close these issues.

2 participants