February 2026 monthly summary for awslabs/mcp: delivered the DynamoDB Cost Calculator Tool enabling cost estimation from access patterns without provisioning infrastructure. The feature supports all 10 DynamoDB operations, computes RCU/WCU and monthly costs using on-demand pricing (us-east-1), and tracks GSI write amplification and storage costs. Input validation is implemented with Pydantic discriminated unions, with actionable error reporting. Generates a Markdown cost report appended to dynamodb_data_model.md. Updated the architect prompt to collect item size, RPS, and item count per access pattern and instructed the LLM to invoke the calculator after completing the data model design. Primary work captured in commit b44905fbd0a9a3772c5e30640444f8e12d6f5f91; co-authored by Lee Hannigan and Sunil Yadav.

Month: 2026-01 — Key features delivered: • DynamoDB Table Provisioning via CDK Tool: A CDK-based app to provision DynamoDB tables from a JSON data model, including validation, Jinja2-based deployment templating, and resource parameters (tableName, partitionKey, sortKey, globalSecondaryIndexes, timeToLiveAttribute, removalPolicy). The CDK app is generated via a new generate_resources MCP tool and is designed to run after dynamodb_data_model_validation. • Enhanced Observability: Implemented exception logging in the DynamoDB MCP server to capture error messages with function context for faster debugging. Major bugs fixed: • Improved error reporting and observability, reducing triage time by ensuring exceptions are logged with actionable context. Overall impact and accomplishments: • Enabled automated, model-driven DynamoDB provisioning, resulting in faster, reproducible deployments and reduced manual steps. • Strengthened reliability through validation-first workflows and enhanced observability, improving incident response and maintainability. • Demonstrated strong IaC practices and tooling across CDK, templating, and MCP workflows. Technologies/skills demonstrated: • CDK (infrastructure as code) and TableV2 constructs; • Jinja2 templating for deployment templates; • JSON data modeling for resource definitions; • MCP tooling and workflow orchestration; • Structured exception logging for improved debugging.