How to structure the project with Kratos?

[spa5k]

Let's say, I have a typical setup, a next js website for the frontend, a separate nodejs based backend using Postgres for data.

So, now I want to add Kratos to this setup, how should i work on adding it optimally?

1. Adding kratos on same database as the one getting used by the main backend.

Pro - no separate database management.
Cons - it mixes up kratos tables onto main schema.

2. Adding kratos on a different database but on same docker container as the one used by main backend.

Pro - not much of a separate database management.
Does not add tables and things to main database.

So what will be the optimal way to access data from kratos in my project.

For first scenario -

1. Since kratos is in same db I can just reference identity id to a kratos id in users table and use it for reference.

For second scenario-

1. Keep a column kratos id in my users table, and then query to the kratos database, probably through the http interface or something, idk.

[vinckr]

It seems like the second scenario has fewer drawbacks.
As you have more practical experience with this setup, would you agree @Benehiko ?

[Benehiko]

Hi @Unbuttun

I would keep the database separate (but you can keep it on the same PostgreSQL instance) between Kratos and your actual application. I would then add an identity UUID column in my application which is basically the identity UUID given by kratos.

You have a couple of options to then populate the columns for a user upon user registration (a new user is created).

• Either you add a middleware that checks the cookie against Kratos with a /sessions/whoami and retrieves all the information and then compare it against your DB (does this user exist?)
• Have an endpoint on your application that can be called by Kratos post-registration hook. See the web-hooks docs

To get the user ID on a request just call the Kratos /sessions/whoami endpoint from your application (middleware).
See the HTTP API docs

[spa5k]

Thanks for the help @Benehiko. 🥂