Let me know how I can improve it -- I might be off in some of the answers. I left a few questions empty/incomplete as I wasn't sure of their answers:

• QHub vs JupyterHub ?
• What technologies does QHub rely on?
• What parts of QHub are not open source?
• Why open source?

Tony, Ani and I discussed that it would be better if the documentation is designed in a way that directs the reader to different pages based on how they identify themselves (a general user -team member etc- or a producer (DevOps person).

• Those with the need to customize qhub setup for a large team (DevOps is the persona we have in mind here) to do so in a more detailed and technical level.

• Those who are simply general users of the tool and don't need to (or have authorization to) make changes in the deployment settings etc.

The landing page for docs will lead general users, which we assume would be a member of a team, to a documentation page that does not contain many of the low-level details about how qhub works in the backend (kubernetes, the entire cloud architecture, etc.).

A producer, on the other hand, is one who needs to maintain such deployment specifics would want to know and be able to customize to serve the needs of an entire team. Thus, their documentation page will entail more low-level technical details and "how to"s.

The documentation addressed to these two different audiences will be captured on two different pages.

Adds a section on spinning up a dask cluster in qhub. This definitely needs modification, but is an initial effort given my limited understanding of qhub. Related to https://github.com/Quansight/qhub-home/issues/26

@ftarlaci Do you know a good place to put this info? When you spin up a local dask cluster, you can access the dask dashboard by adding "/proxy/port/status" to the base url. This method should work for any local urls on the instance running jupyterlab, and is possible due to jupyter-server-proxy. I poked around, but I didn't see a section on spinning up a Dask cluster yet.

Adam Lewis 1 day ago I used the Dask jupyterlab extension to create a cluster, and could see the diagnostics, but when I used it, it created the cluster with the software versions of the base conda environment rather than the environment I wanted to use.

Adam Lewis 1 day ago I was able to figure this out. Since jupyter-server-proxy is already installed, I can access the dashboard url via: https://jupyter.training.qhub.dev/user/[email protected]/proxy/8787/status (edited)

Dharhas Pothina 23 hours ago @ Adam Lewis Lets add this to the qhub docs. Ask @ Fatma a good location for it in te dask section