Scaling issues with writing custom tables and SQL statements

I’ve three queries related to the database and custom SQL operations.

  1. I saw an example using two databases for Nakama in the documentation
nakama --database.address "root@db1:26257" --database.address "root@db2:26257"

I want to understand in what manner Nakama connects to the two databases and uses them.

  1. If we are using custom SQL to create our own tables and perform operations, and assuming we are writing the most optimized query - does it cause issues at the time of scaling when using Heroic Cloud for deployment purposes? I would assume a yes as it may cause issues in how databases are distributed across different instances for scaling the CockroachDB.

  2. Are there any performance or development advantages of using CockroachDB with Nakama instead of Postgres, specific to Nakama? On a general basis, I can understand the advantages and disadvantages of both options, however, I want to understand how it’ll impact our deployment on the Heroic cloud when the database is being handled by Nakama.

Can I please get a response on the same as we are in the critical stage of making a decision for the database engine?

The main intention is that we are using Heroid cloud and we want to make sure that our local environment of the database is the same as Heroic Cloud. As per the documentation, Nakama officially supports CockroachDB, so is it a good assumption that the Heroid Cloud uses CockroachDB for deployments instead of Postgres?