MariaDB Scaling - Databases vs Collections vs Documents

I'm curious about the performance and scaling implications of having many databases as opposed to having many collections with fewer databases or finally the fewest database and collections with documents isolated by an id attribute.

Which complicates horizontal scaling more? Having thousands of databases or thousands of collections inside a single database or finally the fewest databases and collections with a very large number of documents isolated by an id attribute?

The isolated groups of data will have very few dependencies between each other in my project and I want to make heavy use of the real time API.

A database on appwrite isn't a database on mariaDB

hmm interesting, how does that translate to mariaDB?

Only one mariadb database and each collection is a table

If you want to use realtime, build your database around it

The appwrite realtime is one of the most efficient I've tested, especially because there is only one websocket channel. The team did a great job 🙂

Awesome, could you explain a bit more what you mean by building the database around it?

How is appwrite database isolated in mariadb side?

What more are you asking about?

If its a single database in MariaDB side how does appwrite differentiate between collections in different appwrite databases?

We prefix the table name with an internal project id and database id. Feel free to spin up a local instance of Appwrite and connect to the mariadb database to see