We don't know when generation hooks takes more time than expected, massively impacting generation time
We have some generation hooks that happen for all nodes one time, and other that happen for each node. We have a timeout for them, but which is rather hight (5 minutes for a warning, 1 hour for timeout).
This is because we can't know in advance what is a normal time for a hook. Perhaps we could add an header or something in the name to tell rudder what is the expected timeout for that hook.