Running your own services means managing servers, scaling capacity, patching operating systems, and paying for resources even when they sit idle. For many workloads, that is a lot of work that does not directly improve the product.

Serverless architecture hands much of that work to the cloud provider. Your code runs on managed compute and managed backend services. The platform scales them up and down, and you usually pay based on usage.

You stop managing servers directly, but you still own the system design. You still need to think about events, permissions, retries, duplicate messages, cold starts, monitoring, service limits, and cost.

This chapter explains serverless architecture, its core components, how execution works, common patterns, data choices, operational challenges, and when dedicated compute is a better fit.