Serverless seems to be the most cheaper and faster solution to build API: it’s only a piece of code that run in somewhere in the cloud. But what happens if the code isn’t performant, secure, maintainable and if the API isn’t well designed? And what happens if you choose a third part services that could became a bottlenecks of the architecture? Remember moreover that code is always portable, architecture isn’t. It will also demonstrate that DevOps practices help to manage efficiently the lifecycle of a serverless services, and how to have a good diagnostic and monitoring system helps to take the right decisions for operations and development. Good design and monitoring, in general mean to have always the control about costs, and this is still more true with the serverless architectures with which it isn’t possible make reproducible and repeatable the real cloud context in which the services are consumed.