Clean Node.js Architecture: The Service-Repository Pattern

As your Node.js project evolves, it's easy to fall into the trap of cramming all your logic into Express route handlers. However, when your SaaS expands to manage hundreds of endpoints, this "Fat Controller" approach can quickly spiral into a maintenance nightmare. To tackle this challenge, professional engineering teams implement a Service-Repository Pattern, which effectively decouples business logic from both the delivery mechanism (HTTP/API) and the data storage (SQL/NoSQL).

Layer 1: The Controller (Entry Point)

The primary responsibility of the controller is to handle incoming requests. It should validate input schemas (using libraries like Joi or Zod), extract necessary data from req.body or req.params, and invoke methods within the Service layer. Importantly, the controller should never directly interact with the database. If you ever find yourself writing User.findOne() inside a controller, you are deviating from this pattern.

Layer 2: The Service Layer (Business Logic)

Here’s where the magic truly happens. The Service layer is the heart of your business rules. For instance, when a user signs up, the UserService handles critical tasks such as hashing the password, checking if the email is blacklisted, and triggering events for the notification system. By keeping the Service layer decoupled, you gain the flexibility to reuse the registerUser logic across various contexts—be it an API call, a CLI tool, or a Cron job—without rewriting a single line of code.

Layer 3: The Repository Layer (Data Access)

The Repository layer serves as an abstraction for your database interactions. Regardless of whether you’re using Mongoose, Sequelize, or raw SQL, all database queries reside here. This strategic separation allows for seamless transitions between database engines in the future. For example, if you decide to migrate from a local MongoDB setup to AWS DynamoDB, the only changes needed will be within your Repository implementation.

Expert Takeaway: Dependency Injection

To elevate this architecture to a truly "Clean" standard, consider implementing Dependency Injection (DI). By passing the Repository instance into the Service constructor, you can effortlessly mock the database during unit testing. This approach enables you to test complex business logic quickly and efficiently, without needing to connect to a live server.