Hub shows "Spawn failed" message after jupyter servers are culled

I have set up a version of Zero to Jupyterhub on a Kubernetes cluster.

I use the culling options to stop Jupyter servers pods after a certain amount of inactivity, just as is:

    enabled: true
    timeout: 21600
    every: 900

It behaves as it should, the pods are stopped after a while, but when users try to access their Jupyter server that has been automatically stopped, they end to this endpoint /hub/spawn-pending/<username>

It shows this page:

Users “just” have to click on the “Relaunch server” button to trigger the spawn, but I am wondering if there is a way to automatically spawn they arrive on this page.

I also have to say that it only happens the first time users are accessing their server, then it’s automatically relaunched.