advanced

AWS

Amazon cloud services commonly used for compute, containers, serverless, storage, databases, messaging, identity, CDN, and observability.

AWS interviews focus on the services most teams touch first: EC2 or containers, Lambda, S3, RDS or DynamoDB, SQS/SNS, CloudFront, IAM, and CloudWatch. Answers should name trade-offs, not just service acronyms.

Subtopics: EC2, ECS/EKS, Lambda, S3, RDS, DynamoDB, SQS, SNS, CloudFront, IAM, CloudWatch.

On interviews: design a file-upload API with presigned S3 URLs, an SQS worker, RDS connection pooling, and IAM roles—no static keys on instances.

Common pitfalls: public S3 buckets; overly broad `AdministratorAccess` in CI; Lambda without dead-letter queues; RDS without connection limits for Node pools.

The trade-off is AWS ecosystem depth versus operational surface area and bill complexity.

Checklist:

Prefer IAM roles over long-lived keys.
Model async work with queues and idempotent consumers.
Scope S3 policies and encryption explicitly.
Correlate logs, metrics, and alarms for production paths.