Skip to content

Conversation

@ncclementi
Copy link
Contributor

  • Make clear how to setup things via docker compose with rapids notebook, add a few screenshots and a template of a docker compose yaml.

  • Note that I'm only exposing the Jupyter port (8888) to simplify things, but if dask is desired, I think we need to sort how to expose the ports properly because the scheduler one it always shows unhealthy, while the dashboard one turns healthy when creating a Client (this makes sense).

@jacobtomlinson I think we can pair on the scheduler port at some point and improve the docs for this case.

While I'm out push as much as you need to the PR to get this one across the line.

@ncclementi ncclementi requested a review from a team as a code owner October 9, 2025 20:58
Copy link
Member

@jacobtomlinson jacobtomlinson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Overall this looks great!

We can talk about Dask later, but I expect for the most part people will access Dask just jupyter (at localhost) so it's less of a concern.

Brev only provides an HTTP proxy and can't proxy TCP, so if you really did want to get Dask to run through the proxy you would need to use the websocket comm. I expect if you set the port to 8787 and then did LocalCUDACluster(protocol="ws") it would show as healthy.

Co-authored-by: Jacob Tomlinson <jacobtomlinson@users.noreply.github.com>
@jacobtomlinson jacobtomlinson merged commit 8e273b7 into rapidsai:main Oct 15, 2025
3 checks passed
@ncclementi ncclementi deleted the brev-launchable branch October 15, 2025 16:28
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants