I’d love some official guidance on how to integrate: GitHub - jupyter-server/jupyter-scheduler: Run Jupyter notebooks as jobs into a zero 2 k8s installation of JupyterHub! Has anyone done this?
jupyter-scheduler would be a package you install in your user environment, so this would be the relevant docs, where you do:
pip install jupyter-scheduler
to add it to your user image.
If you use the defaults, these will be local jobs, run in the user’s container, and only while the user’s server is running. If you want to hook it up to the cluster to run jobs on kubernetes, you would need a kubernetes implementation of the manager classes, and need to grant the user environments permission to launch kubernetes jobs. I’m not sure if anyone has implemented that, yet.
2 Likes