Matteo Mortari

@tiansiyuan awesome, so it seems a connection issue between your notebook (which seems to be inside a VM, so not sure about that setup...). This makes me believe I should...

@tiansiyuan > By the way, there are something not accurate on https://www.kubeflow.org/docs/components/model-registry/installation/, like: > 1. `You can skip this step if you have already installed Kubeflow >=1.9. Your Kubeflow deployment...

> The latest version of ml-metadata is 1.15.0, it supports python 3.9/3.10/3.11 as mentioned in my previous comments it does not, as this project/reproducer demonstrates: https://github.com/tarilabs/demo20240601-mlmdversions attached screenshot of my...

@tiansiyuan I'm glad to hear you concur about MLMD requirements and that the workaround is applicable to your scenario and that it works. For the "user token" ~~appears to be...

> I still have the timeout issue. That's very strange @tiansiyuan :thinking: looks like a connection issue between the Notebook and the service, on your setup. I've double-checked again, and...

> Any other way I can debug it? > Thanks To me this shows some difference in the KF platform on your setup, or something else might be influencing in...

@tiansiyuan thanks, so all else equal, can't think of something beyond KF at play. Can you share details of which Kubernetes environment are you using? Which version of Minikube (if...

@tiansiyuan I believe on your testing cluster since network policies are enforced by default, you need this: https://github.com/kubeflow/manifests/pull/2724

the 2 tutorials in "getting started" are not necessarily _strongly tied_ together, but fair point I've added a follow-up task into: https://github.com/kubeflow/model-registry/issues/109#issue-2309990930 I'm glad to hear that solved your connection...

as discussed on the KF bi-weekly 2024-06-10, closing this issue as we have at least 2 diff k8s environment tested with KF 1.9 rc (s) (we will open new ones...