What is the most annoying fixed "tax" of a standard AWS infrastructure for an application with asymmetric or low-to-medium traffic? The answer is clear: the fixed costs of always-on compute (ECS Fargate/EC2), relational databases (RDS), and networking (NAT Gateway and Load Balancers). We often end up paying dozens of dollars a month for systems that spend 99% of their time idle, simply because "that's how it's done in production."

In this talk, we will challenge these infrastructure dogmas and demonstrate how to drastically slash operational costs by bringing a standard monolithic Django application into production for under $3 per month.

Debunking the myth that production always requires a dedicated database, we will deploy a SQLite database on a persistent network filesystem (AWS EFS) and use the AWS Lambda Web Adapter to run our standard Django container on a "zero-cost-at-rest" Lambda. We will dissect the entire architecture: from aggressive CloudFront caching to eliminate cold starts for end-users, to managing single-writer concurrency in the admin panel, and handling surgical cache invalidations via Django signals.

No hand-waving or oversimplified tutorials: we will analyze a real-world case study, exploring the pros, cons, consciously accepted architectural trade-offs, and the exact configuration needed for a rock-solid, secure, and sustainable deployment.

Target Audience: Cloud Architects, Tech Leads, DevOps Engineers
Preferred Session Duration: 30 - 35 minutes