Are there time out limits, kernel keeps quitting, is my server too small?

I’ve followed the instructions for TLJH for Azure setup, and sized the server as D16s_v3 which has 16CPU and 64gb, set memory and cpu limits at 2gb and 2cpu, but the kernels are still dying after say 5-10 minutes of inactivity. I’ve searched the docs to see if there’s a time-out setting that checks for inactivity, there must be, sorry I can’t find it, it’s probably totally obvious.

Can anyone tell me how to increase the time-out limits if there are such limits, or if that’s not the problem, help me keep the kernels alive longer. Could my server still be too undersized? Could it be connection issues? Thanks!

Isn’t it here in the documentation?

" Every server found to be idle for more than 10 minutes will be culled."

Also, see this discussion in this Jupyter discourse forum.

Those are the first and second hits Google returns if you search ‘tljh inactivity’.

TLJH works with the free mirco tier on Amazon which is vastly inferior to what you’ve provisioned. Your server is definitely not undersized at this early point in your tests.

1 Like