What is serverless computing in AWS?

Question

Accepted Answer

Serverless doesn't mean 'no servers' — it means AWS manages the servers, scaling, and capacity so you focus only on code and config. Lambda — event-driven functions; scales 0 → millions, pay per ms of execution. Fargate — serverless containers for ECS/EKS; no EC2 to manage. API Gateway — serverless HTTP/REST/WebSocket front door. DynamoDB on-demand — serverless NoSQL; auto-scales read/write. Aurora Serverless v2 — autoscales RDS Aurora capacity by ACU. S3, SQS, SNS, Step Functions — all fully serverless data and integration services. Trade-off: serverless eliminates ops but introduces cold starts (mitigated by SnapStart and Provisioned Concurrency), per-request pricing surprises, and vendor lock-in. Define serverless as 'no servers to manage' — not 'no servers exist.' Mention cold starts and SnapStart to show you know the gotcha and the fix.