AWS reference

Most of this handbook’s AWS material is organised by service — compute, storage, networking, databases, security. This shelf holds the handful of facts that belong to no single service and that are easy to get wrong because they are rarely written down in one place.

Start with global services and single-Region dependencies when designing a failover plan: it sets out which services are not scoped to a Region, where their control planes actually live, and which apparently Regional services carry a hidden dependency on us-east-1. The published service level agreements page records what AWS commits to in writing for the main compute services, and which of them publish nothing at all. What serverless means is worth reading before taking the label at face value on a pricing model. The further reading page collects AWS’s own whitepapers on organising an estate.

For anything service-specific, start from the AWS section index instead.