1Mio users on a InfluxDB server - best practice


Lets say I try to setup a InfluxDB System for 1Mio users.
What is here the best practice?

Visualisation is Grafana.
No User should have access to another user data.

Did you recommend to create for each an own organization?
Or could I use just 1 organization and for each an own bucket?
Or maybe also same organization, same bucket and different measurement?

All steps I need of course with a script. Just a unique name and then my script should create all setups for the new user.

Thank you!