Serverless Architecture Patterns - Event-driven design, cold-start optimization, and cost engineering for FaaS platforms. #1007463

Serverless Architecture Patterns is a practical engineering guide to building event-driven systems on function-as-a-service platforms. It starts from a single observation: the server is still there — the physical machine is running in a data centre owned by a cloud provider — but the operational surface that used to consume an engineer's attention is gone, and that changes what is possible and at what cost.

The book walks through the full discipline — the case for serverless and what it replaces, function-as-a-service fundamentals and the execution model, cold starts and the mitigation strategies that make latency predictable, event-driven design across synchronous, asynchronous, and poll-based event sources, state and storage and the patterns that compensate for statelessness, serverless APIs and the gateway patterns, data pipelines and the engineering that makes them robust, serverless ML and AI inference, observability for stateless systems, cost engineering with the specific levers that produce real savings, security and IAM and the event-source risks, testing and local development workflows, deployment and versioning and traffic shifting, and the trends reshaping the field.

It covers the failure modes that quietly wreck serverless deployments: a recursive invocation that exhausts the account's concurrency limit and takes down every function in the region, a cold start that adds three seconds to the first request after every deploy, an event source mapping that retries a poison message forever and blocks the entire stream, an IAM role that grants wildcard permissions because narrowing them was inconvenient, an alias that points at the wrong version because the deployment pipeline has a bug, a cost spike from a retry storm that no alert catches until the bill arrives. Each is presented with the failure, the countermeasure, and the operational tradeoff.