Ian Bolliger

Actually, while the URL is correctly printed for the local file comparison, it still is not rendering properly in the browser. Typically, in this kubernetes jupyterhub deployment, exposed ports can...

@azawalich unfortunately I was never able to get this to fully work. As for the proxy approach, you're ability to access it is probably controlled by whatever network your jupyterlab/hub...

@vidartf both of those are use cases I've encountered. The overall use case is diffing two notebook files from within a jupyterhub environment.

@vidartf Thanks! Yes I think that would be a good way to compare notebooks. In the interim since my first comment, code-server (browser based VScode) now supports notebook diffs as...

I believe you just need to provide the names of the two branches. For instance, `nbdiff-web branch1_name branch2_name test.ipynb` will show you the diffs using branch1_name as the base. You...

Thanks @consideRatio ! We bumped from helm chart v0.9-4300ff5 to v0.10.6 (sorry, kind of a big jump...). Let me know if there are any other details I can provide

Watching the `traefik` pod, when the scheduler goes down I see errors like: ``` Error during connection: readfrom tcp 10.4.1.18:57274->10.4.0.23:8786: write tcp 10.4.1.18:57274->10.4.0.23:8786: write: connection reset by peer Error while...

Thanks @TomAugspurger that was it! I wasn't even thinking about scheduler size... Do all computed objects pass through the scheduler? I had not adjusted the default scheduler memory which was...

Thanks for the thoughts @TomAugspurger and @quasiben . @quasiben - Just to clarify, the casting from NumPy arrays to dask arrays was just one example of something that broke our...

Just starting to try out this functionality as well. Will be very helpful when we're able to stably deploy a scheduler on a remote pod. Looks like #154 was merged...