Frequent restart/dismiss prompts and NoKernel! problems

Hello all. I’m leading a team that’s using JL quite extensively, with a bespoke extension that acts as a “storefront” to allow us to deploy notebooks to a wide audience. So far this is working extremely well - a credit to all those involved in Jupyter development.

The one fly in the ointment is that we’re starting to see pretty frequent restart/dismiss prompts, presumably due to network instabilities and timeouts in the system somewhere. This is making the experience for my dev team and our users pretty miserable.

The technical details are captured quite nicely in these two Github issues:

It’s clear we’re not the only ones with this issue, and with the current demand for remote work and increased network load, this is becoming more crucial.

@jasongrout mentions in the first of those issues that 2.0 may resolve the issue - has anyone seen this problem resolved in the wild?



I am having the same issue with JupyterLab 1.2.6 and about to upgrade to 2.x. Can anyone confirm that this is resolved in the 2.x version? If so, I’ll skip solving this issue in current version. Thanks!

I am using 2.2.6 and it is having the same issue.
Jupyterlab 2.2.6
Jupyterlab_server 1.2.0

We’re almost 2 weeks after upgrade to 2.x and unfortunately, this issue still exists. Does anyone have some updates on how to deal with it (or at least some experience with tracking the root cause)? Thanks!