A workflow may retry after a timeout.
Make one result for each processing date valid even if the same run executes twice.
Which automation approach is most appropriate?
Use a processing-date key with MERGE or atomic replacement so repeated input converges on one output.
Detailed explanation
Missing or duplicate data is hidden from downstream consumers.
Missing or duplicate data is hidden from downstream consumers.
A cached response does not prove that the data commit completed after a timeout.
A cached response does not prove that the data commit completed after a timeout.
Repeating the same input updates the same logical result instead of appending a duplicate.
Repeating the same input updates the same logical result instead of appending a duplicate.
Each retry adds another row and changes the aggregate.
Each retry adds another row and changes the aggregate.
Check official documentation
Compare the following points with the official documentation.
AWS Step Functions retry、Amazon Redshift MERGE、idempotencyの公式説明を確認する。
What to confirm
タイムアウト後の二重実行を考慮した出力設計を説明できる。
Key points
- Idempotency
- MERGE
- Natural key
- Staging
Notes
- Environment: AWS公式DEA-C01試験ガイドと各サービスの公式ドキュメントの確認
- Confirm current AWS service behavior in the official DEA-C01 documentation.
Foundation review
At-least-once effects
Retries and duplicate deliveries require output-side duplicate absorption.
Commit boundary
Keep unvalidated or partial output out of the consumer-facing table.