Analytics
Amazon Athena
Amazon Athena runs serverless SQL over data in Amazon S3 and federated sources. In GenAI systems it exposes governed analytical context without provisioning a warehouse.
Key points
- Uses the AWS Glue Data Catalog for table metadata
- Supports prepared statements, workgroups, and per-query data limits
- Athena federated query can reach supported external stores through connectors
When to use it
- Query curated S3 data before prompt construction
- Analyze model invocation and evaluation logs with SQL
Exam tips
- Choose Athena for intermittent SQL on a data lake; use a database for low-latency transactional lookup
- Partition and compress S3 data because scanned bytes drive both latency and cost