Long running kernels without browser activity

I am setting up a Zero-to-JupyterHub installation with EKS, and have all the default configurations for the culler as below

cull:
 concurrency: 10
 enabled: true
 every: 600
 maxAge: 0
 removeNamedServers: false
 timeout: 3600
 users: false

However, I want some users to retain their servers beyond the cull timeout incase their kernels are running, even if there is no activity from the browser. Any suggestions on how that could be done?

An ideal solution would be where we could check the individual servers for utilization/kernel activity and decide whether to terminate it accordingly. But Iā€™m not sure how that would work with the hub culler.

1 Like