Background

A fintech startup needed to streamline their manual deployment pipeline and improve release frequency without compromising quality. Their core product stack included:

• Frontend: React (Next.js)
• Backend: Node.js (NestJS)
• Database: PostgreSQL (hosted on AWS RDS)

Objective

• Replace manual FTP deployments with CI/CD
• Introduce environment-based builds (dev, staging, prod)
• Enable rollback functionality in case of deployment failure

Tools Used

• ⚙️ GitHub Actions (CI/CD workflows)
• 🐳 Docker & Docker Compose
• ☁️ AWS EC2 & S3
• 🔐 AWS Secrets Manager

Pipeline Architecture

1. CI Phase:
   • Run tests (Jest + ESLint)
   • Build Docker image tagged by branch and SHA
2. CD Phase:
   • Push to AWS ECR
   • Deploy on EC2 with SSH trigger

Sample GitHub Actions Snippet

jobs:
  build-and-deploy:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout Code
        uses: actions/checkout@v3

      - name: Build Docker Image
        run: docker build -t app:GITHUB_SHA .

      - name: Push to AWS ECR
        run: docker push :GITHUB_SHA
      - name: Deploy to EC2
        uses: appleboy/ssh-action@v0.1.5
        with:
          host:  secrets.EC2_HOST 
          username: ec2-user
          key:secrets.EC2_SSH_KEY 
          script: |
            docker pull :
            docker stop app || true
            docker rm app || true
            docker run -d --name app -p 80:3000 :GITHUB_SHA

Challenges

• Managing secrets across multiple environments
• Handling rollbacks and failed builds gracefully
• Ensuring Docker images didn't bloat over time

Optimizations

• Used multi-stage Docker builds to reduce image size by 60%
• Added Slack notifications for build status via webhooks
• Integrated Git tags to trigger production releases only

Results

Metric	Before	After	Improvement
Deployment Time	45 mins	3.5 mins	92%
Release Frequency	1/month	3/week	12x
Rollback Time	1 day	30 sec	~99%

Conclusion

This transformation allowed the team to ship features faster, ensure higher quality builds, and minimize production downtime. Their entire release pipeline now runs with zero manual intervention.

Next Steps

• Set up blue-green deployments with load balancer switching
• Add SonarQube for code quality reports
• Extend CI pipeline to include end-to-end tests with Playwright