This message was deleted.

Why not use different datasources for each use-case? datasources are pretty isolated.

But then all the data for different applications are in same S3 bucket in the same path right? There is no way then to do access control or stuff like that.