Data AWS Skill

Version: 1.1 Stack: AWS (Glue, Lambda, S3, Redshift, Athena, Step Functions)

AWS data services charge by usage — every query scans data, every DPU-hour costs money, every unpartitioned table means full scans at $5/TB. The difference between a well-architected data lake and a poorly-structured one isn't just performance — it's real dollars. Unpartitioned Athena queries can cost 100x what partitioned queries cost for the same result. Glue jobs at default DPU count waste compute. Lambda functions with clients initialized inside the handler waste cold start time on every invocation.

These patterns aren't just best practices — they're cost controls.

Scope and Boundaries