AWS Cost Analysis

Parses AWS CDK TypeScript/JavaScript projects by traversing the abstract syntax tree to identify all AWS service constructs instantiated in the infrastructure code. Uses static analysis rather than runtime execution to extract service declarations, construct parameters, and resource configurations without requiring deployment or AWS credentials. Maps CDK construct hierarchy to concrete AWS service types (EC2, Lambda, RDS, etc.) for downstream cost analysis.

Fetches and normalizes AWS pricing information from both AWS Pricing API (bulk JSON pricing data) and AWS pricing webpages (HTML scraping for real-time rates). Maintains a unified pricing schema that maps service names, instance types, regions, and pricing dimensions to current rates. Handles pricing updates and regional variations by querying authoritative AWS sources and caching results to minimize API calls.

Maps extracted CDK services to their corresponding AWS pricing dimensions (compute hours, storage GB, data transfer, API calls, etc.) and calculates estimated monthly costs based on resource configurations. Implements service-specific pricing logic (e.g., Lambda pricing by invocations + memory-duration, EC2 by instance-hours + data transfer) and aggregates costs across all services in a stack. Handles regional pricing variations and pricing model selection (on-demand vs reserved vs spot).

Exposes cost analysis capabilities as MCP tools (function definitions) that AI assistants can call via the Model Context Protocol. Defines tool schemas for analyzing CDK projects, retrieving pricing, and calculating costs, with structured input/output contracts. Handles tool invocation from MCP clients, executes analysis pipelines, and returns results in MCP-compliant JSON format. Enables bidirectional context flow where AI assistants can iteratively refine cost analysis based on conversation context.

Automatically discovers CDK project structure by scanning for cdk.json configuration files, tsconfig.json, and stack definition files. Validates project structure against CDK conventions (lib/ directory for constructs, bin/ for entry points) and checks for required dependencies (aws-cdk-lib, constructs). Provides error reporting for misconfigured projects and suggests fixes. Handles monorepo structures with multiple CDK projects.

Caches cost analysis results (service inventory, pricing data, cost calculations) with configurable TTL to avoid redundant computation and API calls. Implements cache invalidation strategies: TTL-based expiration for pricing data (updates hourly), file-based invalidation when CDK source files change, and manual cache clearing. Tracks cache hit/miss rates and provides cache statistics for performance monitoring.

Calculates cost sensitivity to resource parameter changes (e.g., 'what if I double the Lambda memory?' or 'what if I use reserved instances?'). Implements parameterized cost calculations that accept modified resource configurations and compute delta costs. Supports scenario comparison (on-demand vs reserved vs spot pricing) and identifies cost-driving resources. Enables AI assistants to reason about cost trade-offs during infrastructure design.

Analyzes cost differences across AWS regions for the same CDK infrastructure by querying regional pricing variations. Identifies regions with lowest cost and highlights services with significant regional price differences. Generates optimization recommendations (e.g., 'move RDS to us-east-1 to save 15%'). Handles region-specific service availability (some services not available in all regions).

Generates structured cost analysis reports in multiple formats (JSON, CSV, HTML, Markdown) containing cost breakdowns, service inventory, pricing details, and recommendations. Supports custom report templates and filtering (e.g., 'show only services over $100/month'). Exports reports for sharing with stakeholders, archiving, or integration with cost management tools. Includes metadata (analysis timestamp, CDK version, pricing data source).