JupyterLab Kernel Dying Randomly

We have set up JupyterLab in our organization for Data science people. We have this user who reported his Kernel dies at random instances. We have been able to replicate the issue however there’s nothing in the logs, resource usage that suggests what failed or caused the kernel to die. There’s no pattern we could identify even after replicating the issue multiple times, no direct correlation with resource usage or any particular activity. Sometimes the tab is idle; sometimes the user is actively working or running some jobs that the issue appears.


Any help would be appreciated.

Are you importing any specific packages/libraries in each instance? Does it also happen with an empty notebook which had no cells ever executed?

90% of times there was one set of libraries and other times different ones. Regarding empty notebook, we’ll test that scenario as well.