Increase the limit number of active servers on tljh

JupyterHub The Littlest JupyterHub
1

Hello everyone,

I maintain a tljh installation for a university, and we have more and more users on our platform.
Recently, a significant number of simultaneous connections perhaps showed us the limits of tljh: only a hundred users were able to start a server… The others had a negative response from the spawner service (which is normal… tljh is recommended for 0-100 users).
While waiting to deploy a solution more suited to our needs (via swarm or kubernetes), I need to push this limit on tljh.

  • Do you know if this limit of 100 users is strict, or if our problems come from elsewhere?
  • While searching I found the following setting: c.JupyterHub.active_server_limit=0. This setting is supposed to override the active server limit, but it is not mentioned in the tljh documentation. Do you know if this setting can allow me to have a greater number of active servers on tljh?

PS: to do tests, I tried to mass start a large number of user servers from the admin page, but it seems that the servers started by the admin are not subject to a number limit .

Thank you in advance for your answers.

1 Like
2

Yes, this is what you need to set in JupyterHub config to remove that limit. Instead of disabling it (by setting it to 0), maybe you can increase the value to, say 200 (or whatever depending on your demand).

There is another config c.JupyterHub.concurrent_spawn_limit = 100 which limit “concurrent” spawns. So, with such a config, not more than 100 users can spawn single user servers at the same time. This can be disabled as well by setting it to 0.

1 Like
3

Thank you for your answer… I will overwrite this two parameters and hope this will take effect on tljh as well as on Jupyterhub in general.