Hey all!
I’m currently working on pushing a bunch of different applications I have running in the cloud into an Umbrella, mostly so I can manage them as a whole.
One of the things I’ve been mulling over is that I would also like to use the same Postgres instance: at the end of the day it’s not a lot of data and it would make my life infinitely easier. However, since they ARE different applications (for example, a cooking web app and a telegram todo bot), I would like to seggregate the data into different Postgres schemas.
With my current knowledge, how I would do this is configure an Ecto.Repo per application and in each repo set the after_connect property such as: after_connect: {Postgrex, :query!, ["SET search_path TO cooking_app", []]}. This allows me to not have to deal with prefixes on a daily basis and have it isolated there as well as it just simply working out of the box with migrations and everything.
However, the main drawbacks I see to this approach are that (1) as I create more Ecto.Repo instances, I’m creating more connection pools, while I might want to share the same pool between all repos, and also that (2) I still haven’t figured out how to create the schemas easily before running the migrations from each Application.
How would you folks frame this problem? I’m interested in any thought, even if it isn’t necessarily a solution 
Cheers,
Javier
