I made the following change to our extraConfig (z2jh) for the hub’s jupyter_config.py:
c.KubeSpawner.k8s_api_threadpool_workers = c.JupyterHub.concurrent_spawn_limit
The hub comes up fine but I’m not really sure if the change worked or has the value I expect (I expect 64 since that’s what z2jh sets for concurrentSpawnLimit).
I’m wondering if there is a way to dump the config settings for the hub app while it’s running. I poked around the jupyterhub docs and couldn’t really find anything. The troubleshooting docs mention a jupyter troubleshooting command but that doesn’t work [1]. I also don’t see any hub config settings for logging the settings on startup. For example, in OpenStack projects you can configure them to log their config values on startup [2]. That’s quite useful in debugging and making sure your changes are applied correctly. Is there something similar one can do for jupyterhub? If not, are there other tricks like exec’ing into the hub pod and starting a python shell to maybe import the settings (I’m thinking like importing z2jh and using the get_config method).
[1] https://github.com/jupyterhub/jupyterhub/issues/3119
[2] https://zuul.opendev.org/t/openstack/build/6c739619abfb407f81dc8651f6f7e0c4/log/controller/logs/screen-n-cpu.txt#27