Exposing partially KubeSpawner log failure messages

[Renato Alves]

Hi everyone,

I'm trying to figure out if there's any way to provide debugging information in the spawn page to help users diagnose spawn problems themselves. We are using KubeSpawner and currently only a non-descriptive error or timeout is displayed when things go wrong.

The vast majority of spawn failures experienced by our JupyterHub users fall into three groups. We wondered if we could pattern-match those and display something useful to users:
1. Running out of quota or space in their home folder or shared locations
2. Permission errors during spawn when creating specific jupyter files
3. Misconfiguration caused by interfering libraries that get read at spawn (e.g. after a pip install --user ...).

We went through the documentation and saw that events are used by default and this is what is displayed to users, but we couldn't find much about pod log messages. Is there anything we could use to expose the logs entirely or partially?

Thanks,
Renato