Skip to content

community: ensure manifests description is vendor-neutral #4071

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

thesuperzapper
Copy link
Member

@thesuperzapper thesuperzapper commented Mar 30, 2025

Related: #4028
Related: #4014

/area community

What this PR does

This PR proposes an update to the manifests introduction paragraph that clarifies what the manifests actually are.

See diff for the current proposal.

Motivation

@kubeflow/kubeflow-steering-committee I am deeply concerned about a recent change to the manifests description that was merged without wider discussion with the community.

In PR #4028 (about 30 days ago), the manifests description was updated to the following:

The Kubeflow manifests are a collection of community maintained manifests to install Kubeflow in popular Kubernetes clusters such as Kind (locally), Minikube (locally), Rancher, EKS, AKS, GKE. They are aggregated by the Manifests Working Group and are intended to be used by users with Kubernetes knowledge and as the base of packaged distributions.

This wording is misleading, as the manifests are currently only tested to deploy a minimum-viable Kubeflow Platform on Kind, with no guarantees for other platforms.

It's also important to note that the Manifests WG charter explicitly prohibits the inclusion of (kubernetes) distribution-specific manifests, so even mentioning specific distributions verges on violating the charter.

@google-oss-prow google-oss-prow bot added the area/community AREA: Community Docs label Mar 30, 2025
Copy link

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign 8bitmp3 for approval. For more information see the Kubernetes Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@@ -339,9 +339,8 @@ The following table lists distributions which are <em>maintained</em> by their r

### Kubeflow Manifests

The Kubeflow manifests are a collection of community maintained manifests to install Kubeflow in popular Kubernetes clusters such as Kind (locally), Minikube (locally), Rancher, EKS, AKS, GKE.
Copy link
Contributor

Choose a reason for hiding this comment

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

I would prefer not to change this line as it gives an idea to new users that they can get use the manifest to deploy kubeflow on any of the popular Kubernetes platforms available. The change below highlights only kind clusters which may miscommunicate the information that the manifests work only on kind. @juliusvonkohout any comments

Copy link
Member Author

Choose a reason for hiding this comment

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

Up until very recently (30 days ago) no reference to and specific Kubernetes distribution was included, and this did not confuse anyone.

However, the current wording is misleading, as the manifests are only tested to deploy a minimum-viable Kubeflow Platform on Kind, with no guarantees for other platforms.

It's also important to note that the Manifests WG charter explicitly prohibits the inclusion of (kubernetes) distribution-specific manifests, so even mentioning specific distributions verges on violating the charter.

Copy link
Contributor

Choose a reason for hiding this comment

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

Up until very recently (30 days ago) no reference to and specific Kubernetes distribution was included, and this did not confuse anyone.

However, the current wording is misleading, as the manifests are only tested to deploy a minimum-viable Kubeflow Platform on Kind, with no guarantees for other platforms.

It's also important to note that the Manifests WG charter explicitly prohibits the inclusion of (kubernetes) distribution-specific manifests, so even mentioning specific distributions verges on violating the charter.

@thesuperzapper my comment is more about the wording that you are using i.e. The Kubeflow Manifests are community maintained kustomize manifests which are tested to deploy a minimum-viable Kubeflow Platform on **[Kind](https://kind.sigs.k8s.io/)** clusters. while the old one (30 days ago) only had the Kubeflow manifests are a collection of community maintained manifests. The change you are proposing kind of throws users that the manifests will only work on kind which is what I am asking you to correct. The previous one explicitly dind't mention only kind and hence it never sounded confusing to new users like me. But the one you are proposing is what i find a bit confusing/misleading and hence the ask to update it to match the older one to avoid any confusions.

Copy link
Member Author

Choose a reason for hiding this comment

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

We could revert the change from 30 days ago, and not reference any specific distribution. However, it's important for people to understand that we only officially test the kubeflow/manifests on Kind, and make no guarantees about any other Kubernetes environment.

There is also a governance requirement (from the Manifests WG charter) to ensure that no specific distribution gets preference in the kubeflow/manifests, hence why we only test Kind. This is critical for the health of the project, and ensuring we remain vendor neutral.

PS: I want to stress that as always, anyone is welcome to make distributions for specific platforms (and many have), but those are not official, nor can they live in kubeflow/manifests.

Copy link
Contributor

Choose a reason for hiding this comment

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

I agree its better to highlight that they have been tested to work only on kind and not on other kubernetes platforms

Copy link
Member

Choose a reason for hiding this comment

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

We are vendor neutral, we just work fine on the most Kubernetes clusters.
So I explicitly want to keep the existing message. We have many users who succesfully install from Kubeflow manifest on these types of Kubernetes clusters.

@vikas-saxena02
Copy link
Contributor

@thesuperzapper I just put a comment on the proposed change. The existing wording looks better to me.

@juliusvonkohout
Copy link
Member

juliusvonkohout commented Apr 1, 2025

We discussed changes for a long time in the KSC and we will continue to get exactly those questions and users raising issues. This paragraph covers many questions that we typically get, so it is very useful.

Even if we change it to
"The Kubeflow Manifests are a collection of community-maintained manifests for installing Kubeflow on popular Kubernetes clusters in the cloud an on-premises" it will be in practice the same. Users will install it on the popular Kubernetes clusters. Andrey even did a mini survey with the audience at Kubecon, which strongly supports that statement. We are a best-effort open source project and that is also why the KSC has shortened the text so much. There is no need to present us worse than we actually are and discourage users.

Also the updated diagram in https://github.com/kubeflow/website/pull/4038/files shows that we as Kubeflow try to run on that infrastructure.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
area/community AREA: Community Docs size/XS
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants