jupyterhub

Contributed recipes are not checked in CI, were mostly added a long time ago and so they might be out-of-date.
It would be nice if someone checked them, updated packages versions used there and verified that they still work.

Following up on jupyterhub/binderhub#1017 (comment). In particular the comment about "reduce duplication".

A possible way to do that which I've seen in other projects is to add an endpoint to the backend from which the UI loads the config. We'd have an endpoint like binder.example.com/_config which returns a JSON which is then used to configure the dropdow

Follow up tasks from #908:

• work out what a "gitpod link" that opens the editor in a documentation file (not README.md) or a file that welcomes people (some instructions) looks like
• mention gitpod as an option for contributing to the docs in CONTRIBUTING
• add a badge to CONTRIBUTING
  • they have a badge like mybinder.org does and I am sure it is somewhere in https://www.gitpod.io/docs/ b

Proposed change

I'd like to see our autohttps deployment be able to provide metrics that prometheus can scrape.

Who would use this feature?

Users of the helm chart that wants to ensure they don't bottleneck because low resources on this proxy pod.

(Optional): Suggest a solution

I suggest we experiment with [this Traefik configuration](https://docs.traefik.io/observability/metr

When a user has a version of gradebook.db that has not been upgraded to the latest version,
Formgrader will hang when it tries to open the database file.
The error is only visible in the log output and not propagated to the web interface.

Originally reported by @ReubenTheDS in jupyterhub/the-littlest-jupyterhub#638.

Proposed change

Add some info in the docs about the authenti

I think it would be better to add stop_timeout instead of re-using start_timeout here:

https://github.com/jupyterhub/kubespawner/blob/5040e75022f302b3cbfadaa1ae18a5ab9d5563a0/kubespawner/spawner.py#L1784

Per the discussion in #529, the /widgets page should be represented on the homepage. How so is open to debate. One idea would be mimic the module style now used for JupyterLab and Jupyter Notebook.

In a recent debugging session, @minrk pointed out that if you only partially pin your repository (e.g. pin numpy versions but don't pin Python), you are likely going to break future-reproducibility. This is because the non-pinned version may stop supporting your pinned version.

We have a [short guide to reproducib