No longer able to select kernel for Jupyter notebook

[Mjboothaus]

Bug description

I have a environment defined in .gitpod.yml file that creates a conda Python environment (including Jupyter notebooks). It has been working fine for a while now.

Steps to reproduce

When I fired up Gitpod today (from https://github.com/DataBooth/db-hello-gitpod) I can no longer select a Jupyter kernel.

This is the workspace instance - https://green-bass-o31gml56.ws-us15.gitpod.io

I tried stopping and starting the workspace and the behaviour is reproducible.

Expected behavior

When you click Select Kernel within the notebook interface within VS Code you normally can choose from a list of (Python) environments -- but there are now no choices.

Example repository

https://github.com/DataBooth/db-hello-gitpod

Anything else?

Really loving Gitpod.

[Mjboothaus]

So trying to work out what has changed and broken this (as there were no changes in the .gitpod.yml file).

Here are the Editor: Help -> About details:

Version: 1.59.0
Commit: 330c9b35c2f3d95af9192bc1ce23eccc5fcbe808
Date: 2021-08-30T07:53:57.621Z
Browser: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/14.1.2 Safari/605.1.15

[Mjboothaus]

So I assume it is something in the Gitpod (custom) VS Code editor build that has broken this functionality. I will try opening the notebooks via jupyter notebook in the command line as a work-a-round.

[Mjboothaus]

So work-a-round not working at least initially as the notebook via browser is not able to find the kernel.

This might be a Jupyter config issue in this setup - although it is puzzling.

[Mjboothaus]

So this is certainly a blocker -- I'm wondering where I can see release details for Gitpod releases, VS Code releases etc. in a centralised way so that it is easier to pinpoint what part of the chain has been broken.

[Mjboothaus]

Hmmm... still trying to work-a-round:

From Command Palette I tried to Select Interpreter to Start Jupyter server

without success. I also tried to re-install the kernel via:

python -m ipykernel install --user --name=ENV_NAME

which resulted in:

Installed kernelspec ENV_NAME in /home/gitpod/.local/share/jupyter/kernels/ENV_NAME where ENV_NAME is the label I use for my conda environment. But this doesn't seem to have resolved it.

[Mjboothaus]

Maybe this is related to these types of issues:

microsoft/vscode-jupyter#513

microsoft/vscode-jupyter#2282

[csweichel]

/schedule

[roboquat]

@csweichel: We have more than 16 cards scheduled. Above 0 only scheduler can perform this action. Those are:

In response to this:

/schedule

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[csweichel]

/schedule

[roboquat]

@csweichel: We have more than 16 cards scheduled. Above 0 only scheduler can perform this action. Those are:

In response to this:

/schedule

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.