How to do port forwarding on jupyterlab instance?

[shauryagoel0]

I want to run a service on jupyter hub instance and want to expose a port for that so that I can hit it using my PC, like opening the service's dashboard on my browser. I want to give this feature to a Data science team member, and they would not have access to K8s, only tools in nebari like jupyter hub, conda store.

Logically, I believe we would need to open a port on a K8s service (like nebari-jupyterhub-ssh). I could then port forward it on my PC using SSH. But, how can I actually do this?

I tried doing ssh as described above, but added the -L 8000:localhost:8000 flag as well. This didn't work as I believe, 8000 port is not exposed in nebari.

[viniciusdc]

Hey @shauryagoel0, based on your description, it seems like jhub-apps could be the best use case for this. It lets you spin up apps/dashboards on demand and share across teams on Nebari. https://www.nebari.dev/docs/how-tos/jhub-app-launcher/

The above has info on how to set that on nebari, and the following one has info on how to create an app:
https://jhub-apps.nebari.dev/docs/create-apps/general-app

[Adam-D-Lewis]

@shauryagoel0 I agree with Vini that jhub-apps is likely the best based on your description, but another possible solution isjupyter-server-proxy, which is installed on the jupyterlab images. It allows you to access a locally running service listening on a local port, but it's still only accessible behind your user's authentication. If you want to try it out open a jupyterlab instance, run python -m http.server in the terminal and then navigate to https://<your-nebari-domain>/user/<your-username>/proxy/8000/ and you'll see the exposed http server. As mentioned your user credentials are still needed in order to view the running service.

[shauryagoel0]

Thank you. This is what I was looking for.

[Adam-D-Lewis]

Closing for now, but feel free to respond if additional guidance is needed.