You'll want to deploy the backend and frontend each in their own service. You can do this by:

1. Creating 2 services

2. Adding the repo to both

3. For each service, specify the directory the service will deploy from (maybe in your case it'd be /backend and /frontend)

There's a full guide for this here: https://docs.railway.com/tutorials/deploying-a-monorepo

NOTE: This is assuming this is an isolated monorepo with no shared directories or anything, lmk if this isn't the case

Hey rol-xm,

This is a very common point of confusion when deploying monorepos with railway.json, and it's a great question. Your approach of using this file is excellent for keeping your deployment configuration as code (Infrastructure as Code).

The moderator's suggestion of setting the root directory for each service in the UI is a valid workaround, but let's solve your actual question: "Why is my railway.json file being ignored?"

The Reason: The Project Creation Flow

The most likely reason your railway.json is being ignored is because of how the project is being created. If you simply connect a GitHub repository and let Railway auto-detect its contents, it will use its "Railpack" build system by default and completely bypass the railway.json file.

To use your configuration file, you must explicitly tell Railway to do so during the initial setup.

The Solution: Correctly Deploying from railway.json

Here are the exact steps to get it working. You will need to start with a fresh project.

1. Delete the existing, failing project from your Railway dashboard.

2. Click New Project and then select Deploy from GitHub repo.

3. Choose your mbconnect-billing-system repository.

4. This is the crucial step. After Railway analyzes your repo, it will present deployment options. Do not click the default "Deploy Now" button. Instead, look for and select the option that says "Deploy from railway.json file".

5. Railway will now read your railway.json file. Instead of creating one failing service, it will automatically provision and configure the two services ("backend" and "frontend") that you have defined.

Verify Your railway.json File

To be certain your configuration is correct, here is a properly formatted railway.json for a monorepo structure like the one in your repository. You can compare this against your own file.

JSON

{
  "$schema": "https://railway.app/railway.schema.json",
  "services": [
    {
      "name": "backend",
      "build": {
        "builder": "DOCKERFILE",
        "dockerfilePath": "backend/Dockerfile",
        "dockerContext": "backend"
      }
    },
    {
      "name": "frontend",
      "build": {
        "builder": "DOCKERFILE",
        "dockerfilePath": "frontend/Dockerfile",
        "dockerContext": "frontend"
      }
    }
  ]
}

Important Note: Notice the dockerContext property. This tells Docker to use the backend or frontend directory as its starting point for the build. This is essential for commands like COPY . . inside your Dockerfiles to work correctly.

Following these steps should solve the issue and get your monorepo deploying just as you intended.

Hope this helps!