GuideBeginner

Aws Serverless And Containers

This guide continues directly from the AWS Core Services Guide, using the same Andes Cargo case and the same LocalStack lab, to teach the two real ways compute gets designed in production AWS: serverless in depth and containers. The Lambda function that processed shipment manifests becomes a production-grade function with layers, concurrency, and secrets; it gets wrapped in a Step Functions workflow with explicit retries and error handling; the direct S3 trigger gets decoupled with an EventBridge event bus, complete with its own dead-letter queue; and a REST API Gateway unifies HTTP access to everything Andes Cargo exposes. On that same thread, the guide builds `andes-cargo-status-api`, an always-on service (unlike a function that only runs on events), packaged in Docker, pushed to a registry, and run with ECS's cluster/task-definition/service model on Fargate. The guide is upfront about a real limitation: LocalStack's free plan doesn't include ECR or ECS, so those two modules execute the real mechanics with plain Docker and a local registry, and show verified command syntax and output as reference, leaving LocalStack's paid plan as an optional, never-required path. The capstone runs the event-driven flow end to end and documents the container path, closing with the "when Lambda, when a container" criterion installed since the first module.