aws-samples/well-architected-iac-analyzer

What it solves

This project provides a way to automatically evaluate Infrastructure as Code (IaC) templates and architecture diagrams against AWS Well-Architected best practices. It helps users identify where their cloud infrastructure deviates from established standards and provides actionable suggestions for improvement.

How it works

The application uses Amazon Bedrock and generative AI (specifically Anthropic Claude models) to analyze uploaded files. It leverages a Bedrock knowledge base—synchronized with AWS Well-Architected whitepapers—to perform RAG-based analysis. The system can process CloudFormation, Terraform, and AWS CDK templates, as well as PDF documentation and architecture diagrams (PNG/JPG). For diagrams, it can even generate corresponding IaC templates. It also includes an interactive chatbot assistant for personalized guidance.

Who it’s for

Cloud architects, DevOps engineers, and security professionals who need to ensure their AWS infrastructure is compliant with Well-Architected best practices across various lenses (e.g., Serverless, IoT, Machine Learning).

Highlights

• Multi-Format Analysis: Supports CloudFormation, Terraform, CDK, PDFs, and architecture diagrams.
• Prioritization Framework: Uses an Eisenhower Matrix to plot best practices by risk criticality and implementation effort.
• High-Performance AI: Supports latest Claude models with 1M token context windows and adaptive thinking for complex reasoning.
• Parallel Processing: Configurable batch sizes allow for analysis speeds up to 80% faster than sequential processing.
• Custom Lenses: Allows organizations to define their own internal security policies and compliance requirements as custom analysis lenses.
• Cost-Optimized Storage: Uses S3 Vectors as the default vector store to reduce costs by up to 80% compared to OpenSearch Serverless.