Skip to content

Open Codelabs AWS (Amazon Web Services) Deployment Guide

This document describes several ways to deploy Open Codelabs on AWS infrastructure.

AWS deployment options

Depending on your architecture, there are three primary approaches.

Similar to Google Cloud Run, App Runner is a fully managed service that deploys directly from source or container images. It is the easiest way to run both frontend and backend without managing infrastructure.

Architecture

  • Frontend: SvelteKit -> App Runner service 1
  • Backend: Rust Axum -> App Runner service 2
  • Database:
  • Simple: SQLite inside the App Runner instance (data resets on restart)
  • Recommended: AWS RDS (PostgreSQL/MySQL)

Deployment steps

  1. Create ECR repositories 
    aws ecr create-repository --repository-name open-codelabs/backend
aws ecr create-repository --repository-name open-codelabs/frontend
  2. Build and push images 
    # Backend
cd backend
docker build -t [ACCOUNT_ID].dkr.ecr.[REGION].amazonaws.com/open-codelabs/backend:latest .
docker push [ACCOUNT_ID].dkr.ecr.[REGION].amazonaws.com/open-codelabs/backend:latest

# Frontend
cd ../frontend
docker build -t [ACCOUNT_ID].dkr.ecr.[REGION].amazonaws.com/open-codelabs/frontend:latest .
docker push [ACCOUNT_ID].dkr.ecr.[REGION].amazonaws.com/open-codelabs/frontend:latest
  3. Create App Runner services
    • In the AWS Console, open App Runner and click "Create service".
    • Select the images from ECR and configure environment variables.
    • Backend: DATABASE_URL, ADMIN_ID, ADMIN_PW
    • Frontend: VITE_API_URL (backend App Runner URL)

2. AWS EC2 (traditional)

Deploy with Docker Compose on a virtual machine. This can be the cheapest option (especially with Free Tier).

  1. Create an EC2 instance (Amazon Linux 2023 or Ubuntu recommended)
  2. Configure security groups: open ports 80, 443 (HTTPS), 5173 (frontend), 8080 (backend)
  3. Install Docker and Docker Compose
  4. Clone and run 
    git clone https://github.com/JAICHANGPARK/open-codelabs.git
cd open-codelabs
docker compose up -d --build

3. AWS Amplify (frontend only)

Use this if you are running Firebase mode (VITE_USE_FIREBASE=true) or only want to host the frontend on AWS.

  1. Connect the GitHub repo in the Amplify Console
  2. Configure the build settings (SvelteKit is auto-detected)
  3. Add env vars such as VITE_USE_FIREBASE and VITE_FIREBASE_API_KEY

Database and storage (production)

For production, use persistent services for data durability.

1. Database (RDS)

Set the backend DATABASE_URL to your RDS instance. - PostgreSQL: postgres://user:password@host:port/dbname - MySQL: mysql://user:password@host:port/dbname

2. Storage (S3)

You can use S3 for uploads. - The project currently supports local filesystem or Firebase Storage by default. - For S3, you will need to implement an additional SDK integration. A local volume mount or Firebase mode is recommended for container deployments.

Which option should you choose?

Feature App Runner EC2 (Docker) Amplify (Frontend)
Difficulty Low Medium Very low
Cost Usage based Fixed (instance) Usage based
Ops overhead Very low Medium Low
Best for Quick prototyping, auto scaling Lowest cost, full control Firebase mode

Conclusion: If you are comfortable with containers, choose App Runner. If cost is the priority, EC2 is a good fit.