Installing Kubeflow

Deployment options for Kubeflow

This guide describes how to install Kubeflow subprojects, Kubeflow Community Distribution, or vendor-packaged Kubeflow Distributions.

Read the introduction guide to learn more about Kubeflow concepts.

Installation Methods

You can install Kubeflow using one of these methods:

Kubeflow Subprojects

Kubeflow subprojects in the Kubeflow landscape can be deployed as a standalone services, without the need to install the entire Kubeflow distribution. You might integrate these services as part of your existing AI platform or use them independently.

These projects are a quick and easy method to get started with the Kubeflow. They provide flexibility to users who may not require the capabilities of a full Kubeflow distribution.

The following tables list Kubeflow subprojects grouped by their corresponding maturity levels. Learn more about maturity levels expectations and requirements in this document

You can find the list of archived projects in this document.

Graduated Projects

These projects are stable and ready for general availability. Breaking changes are only allowed following the defined feature lifecycle for the project.

Kubeflow SubprojectRoadmap
Kubeflow HubRoadmap
Kubeflow KatibRoadmap
Kubeflow NotebooksProgressing
Kubeflow PipelinesRoadmap
Kubeflow Spark OperatorRoadmap
Kubeflow TrainerRoadmap

Incubating Projects

These projects are actively developed, broadly usable, and on track for Graduation. While most core functionality is stable, it is still maturing toward a final release.

Kubeflow SubprojectRoadmap
Kubeflow KaleProgressing
Kubeflow SDKRoadmap

Experimental Projects

Not all pieces of these projects are in place, and it may not be ready for wider adoption. User feedback around the UX of these projects is desired, such as for Custom Resource Definition APIs, technical implementation details, and planned use-cases for the projects.

Kubeflow SubprojectRoadmap
Kubeflow MLflow IntegrationProgressing
Kubeflow MCP ServerProgressing
Kubeflow MCP Spark History ServerProgressing

Deprecated Projects

Development of this project is halted and no new versions are planned. New issues will likely not be worked on except for critical security issues. Projects assets that are included in the releases are expected to exist for at least two minor releases or one year, whichever happens later.

Currently, Kubeflow doesn’t have any deprecated projects.

Kubeflow Distributions

You can use one of the following methods to install the Kubeflow Distributions.

Kubeflow Community Distribution

See the definition for Kubeflow Community Distribution in the overview page

If you want a stable / conservative experience we recommend to use the latest stable release:

You can also install the master branch of kubeflow/community-distribution by following the instructions here and provide us feedback.

Packaged Distributions

See the definition for Kubeflow Distributions in the overview page

The following table lists distributions which are maintained by their respective maintainers:

Maintainer
Distribution Name
Kubeflow VersionTarget PlatformLink
Canonical
Charmed Kubeflow
[release notes]MultipleWebsite
prokube.ai
prokube MLOps
Multiple Kubernetes distributionsWebsite
Microsoft Azure [release notes]Azure Kubernetes Service (AKS)Website
NutanixNutanix Kubernetes PlatformWebsite
QBO GPU Cloud [release notes]QBO Kubernetes Engine (QKE)Website
Red Hat
Open Data Hub
OpenShiftWebsite

Next steps

Feedback

Was this page helpful?