Infinite loop with print crashes jupyter lab browser UI

[lsignac]

Not sure... but it seems to me that last year, an infinite loop like this :

whie True:
    print("OK")

used to "raise" this error (not a Python exception, a notebook error) :

IOPub data rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.

With the new jupyterlab version available on nbhosting, there is no such error, and the ui freezes before anyone hits the stop button of the UI. This is not a nnbhosting, but a jupyter issue and may be it is related to this issue : jupyter/notebook#2364
Also, may be setting --NotebookApp.iopub_data_rate_limit could prevent the browser from freezing ?

[parmentelat]

hey
the issue you are referring to is about the classic notebook - plus, it's 7 years old
so I don't believe it is relevant...

regardless; I suggest you try this on your end on a regular jupyter lab so that you can freely play with any setting, and once you have a definite answer I can adapt nbhosting
another option of course would be to allow users to set their own setting somehow; however this seems a little overkill at this point I believe, what do you think ?

[lsignac]

I ask students to be more careful with infinite loops :).
I will try settings on my local installation and if an interesting setting appears, I will tell you.

[lsignac]

On my own jupyter lab, the default value for data rate is OK (1 000 000 bytes / sec):

On jupyterlab's nbhosting, the behaviour is different :

With my students, I use the nbhosting classic notebook interface. The same :

Without the rate limitation, an infinite loop with a print inside makes the browser freeze, and it's sometimes impossible to press the STOP button.

Is it possible to set or lower the max data rate ?

On my own machine, I can change the jupyter lab max data rate this way :

jupyter lab --ServerApp.iopub_data_rate_limit=1000000

May be (not tested), for nbhosting classic notebook, it's :

jupyter notebook --NotebookApp.iopub_data_rate_limit=1000000

Thanks.

PS : My jupyter lab version : 4.3.1