jupyter_ai extension ray memory management

[messiah1030k]

Problem

When using the jupyter_ai extension, the memory is used by default around 2GB. (jupyterlab docker image in k8s)
Disabling jupyter_ai extension still does not reduce memory usage. I checked and the ray process is still Running.
Of course, starting jupyterlab in a disabled state reduces memory usage, but I want to reduce memory usage without restarting.

Proposed Solution

1. Provides a way to set the amount of memory allocated to the ray process
2. Disabling the jupyter_ai extension also stops the ray process.

Additional context

[JasonWeill]

Closing this because we have migrated from Ray to Dask (#244). If you continue to notice performance issues with the new system, please open a new issue. Thank you for your contribution!