kubeflow/manifests 1.9.0-rc.2 on GitHub

Third RC of the KF 1.9 release 🎉

The release includes the following platform/manifests/security changes:

Kubernetes 1.27 - 1.29 support, but 1.30+ should work as well. So administrators can run it on supported and secure Kubernetes clusters.
Kustomize 5.2.1+ support with way less warnings, since we modernized most manifests. So platform engineers have a modern and supported installation tool chain.
Oauth2-proxy replaces oidc-authservice, which brings improved token based authentication. ML engineers can now use tokens instead of insecure passwords for automation of Kubeflow from e.g. GitHub actions.
Network policies are enabled for the Kubeflow core services as a second layer of defense before Istio authorization policies. This gives administrators a better network overview and segmentation while also enforcing common enterprise security guidelines.
The Documentation has been overhauled. It now contains guidelines for upgrading and extending the Kubeflow Platform for administrators and enables new users to have a basic installation in a few minutes on their Laptop.
Automatic CVE scanning with Trivy on the master branch is relevant for DevSecOps and enterprise security guidelines
We have updated the third-party components to provide integration with the larger ML ecosystem. E.g. ray is heavily used to train LLMs
- Bentoml 1.2.28 and 1.1.21
- Seldon 1.18.1
- Ray 2.23 and Kuberay 1.1.1
Update of the common components such that we support the latest Kubernetes versions and security patches for platform engineers.
- Istio 1.22.1
- Dex 2.39.1
- Cert-manager 1.14.5
- Knative 1.12.4

We also updated the following components:

You can see each component release notes in the following URLs:

TODO rebuild kserve models web application, kserve/models-web-app#81, kubeflow/kubeflow 1.9.0-rc.1

kubeflow/manifests 1.9.0-rc.2 Kubeflow 1.9.0-rc.2 on GitHub