Deploying Scalable Flask Applications on AWS with GitHub CI/CD Pipelines

Deploying a Flask Application On AWS with CI/CD Pipeline

Lmprojectdevto developed a movie quiz application utilizing the Python Flask framework to interface with a PostgreSQL backend. The system leverages Gunicorn as a WSGI server to dynamically scale worker processes based on the number of logical CPUs available on the host.

Why This Matters

Production-grade AWS deployments require moving beyond built-in development servers to a robust architecture featuring Nginx as a reverse proxy and RDS for managed persistence. Implementing a CI/CD pipeline using GitHub Actions and ECR ensures that containerized images are consistently updated across EC2 instances, preventing the configuration drift and manual deployment bottlenecks that often plague late-stage development cycles.

Key Insights

• Gunicorn handles concurrent requests by dynamically identifying the number of worker processes to generate based on available logical CPUs.
• GitHub Actions automates the CI/CD pipeline by managing AWS secrets, logging into Amazon ECR, and building Docker images for deployment.
• Nginx serves as a reverse proxy on EC2, utilizing Certbot to automate the installation and renewal of SSL/TLS certificates for HTTPS.
• AWS RDS (db.t4g.micro) automates administrative operations such as patching and backups for the PostgreSQL production database.
• Route 53 facilitates domain routing by mapping guessthatmovie.today to the specific Elastic IP address assigned to the EC2 instance.

Practical Applications

• Use Case: Deploying low-traffic web applications on t2.micro EC2 instances to optimize cost-efficiency while maintaining dedicated compute. Pitfall: Failing to implement an IAM Role for EC2, which prevents the instance from securely pulling container images from ECR.
• Use Case: Integrating Redis cache within the Flask application architecture. Pitfall: High volumes of direct GET requests to the RDS instance without a caching layer, leading to increased latency and database load.

References:

• https://dev.to/lmprojectdevto/deploying-a-flask-application-on-aws-with-cicd-pipeline-4i2l