← Back to Digital Garden
Elastic Beanstalk Aurora Serverless ECR API Gateway Docker

Containerized App Migration to AWS Managed Services

📌 Project Description

Manually orchestrating Docker containers on bare virtual instances (IaaS) often results in significant operational overhead. This project demonstrates the architectural migration from self-managed containers to AWS Platform as a Service (PaaS) and Serverless solutions to minimize maintenance burdens and automate scalability.

The core focus of this project is deploying a containerized Node.js application using AWS Elastic Beanstalk, migrating the relational data tier to Amazon Aurora Serverless, and integrating the backend services with a static frontend utilizing Amazon API Gateway.

🛠️ Tech Stack & AWS Services

  • Compute & Container Management: AWS Elastic Beanstalk, Amazon Elastic Container Registry (ECR), Docker.
  • Database: Amazon Aurora Serverless v2 (MySQL-Compatible).
  • Networking & API: Amazon API Gateway, Amazon VPC.
  • Concepts: PaaS (Platform as a Service), Serverless Databases, Stateful vs Stateless Architecture, Microservices Integration.

🏢 Business Scenario

A web application (Coffee Suppliers App) was initially running entirely within Docker containers, including its MySQL database. However, running a relational database (which is inherently stateful) inside a dynamic container poses high risks to data persistence and reliability. To prepare for business growth, the architecture was re-engineered:

  1. Database Layer: Migrated to Amazon Aurora Serverless to seamlessly scale compute capacity up or down based on transaction load, optimizing costs during idle periods.
  2. Web Tier: The stateless Node.js container was deployed to AWS Elastic Beanstalk, offloading load balancing and auto-scaling management entirely to AWS.
  3. Frontend Layer: Hosted securely on Amazon S3, configured to communicate with the backend via a managed Amazon API Gateway.

🚀 Implementation Steps

Phase 1: Serverless Database Migration (Aurora Serverless)

  • Evaluated network topology to ensure the presence of subnets across multiple Availability Zones for High Availability.
  • Provisioned an Amazon Aurora Serverless v2 (MySQL-compatible) cluster to replace the legacy containerized database.
  • Hardened Security Groups to guarantee the database only accepts internal connections specifically from the application instances (port 3306).
  • Interfaced with the Amazon RDS Query Editor and MySQL CLI to instantiate the schema (suppliers, beans tables) and execute the initial data seeding process.

Phase 2: PaaS Container Deployment via Elastic Beanstalk

  • Inspected and validated the Docker image of the Node.js application securely hosted within Amazon ECR.
  • Configured the necessary IAM Role (aws-elasticbeanstalk-ec2-role) to grant the Elastic Beanstalk environment the authorization required to pull images from the private ECR repository.
  • Automated the Elastic Beanstalk environment provisioning via the AWS CLI, injecting critical environment variables (such as APP_DB_HOST pointing to the Aurora Serverless endpoint).
  • Executed the deployment by crafting a Dockerrun.aws.json file, instructing Elastic Beanstalk to proxy incoming HTTP traffic from the native Load Balancer to the internal container port.

AWS Containerized App Migration 1

Phase 3: API Gateway Proxy Integration

  • Configured Amazon API Gateway to act as the intermediary layer between the frontend application (S3) and the Elastic Beanstalk backend.
  • Provisioned a new resource (/bean_products) and established a GET method utilizing HTTP Proxy Integration.
  • Pointed the proxy endpoint URL to the Elastic Beanstalk domain to seamlessly route JSON payloads.
  • Enabled CORS (Cross-Origin Resource Sharing) and deployed the API to the production stage.

AWS Containerized App Migration 2

🎯 Results & Key Takeaways

  • Architectural Modernization (Stateful vs Stateless): Demonstrated advanced architectural understanding by decoupling the stateless compute workloads (Docker/Elastic Beanstalk) from the stateful data storage (Aurora Serverless).
  • Operational Overhead Reduction (NoOps): By leveraging Elastic Beanstalk and Aurora Serverless, the engineering team is relieved from manual OS patching, container orchestration, and database capacity planning.
  • Microservices Integration: Successfully stitched together distinct managed services (S3 Frontend, API Gateway, Beanstalk Backend, Aurora DB) into a cohesive, secure, and highly scalable application ecosystem.