Run an S3 ingestion step, a Glue transformation, and a Redshift load in sequence.
Do not run later steps after a failure and make failed work safe to retry.
Which automation approach is most appropriate?
Define each step as a Step Functions Task and express dependencies, Retry/Catch behavior, and inputs and outputs explicitly.
Detailed explanation
A large retry unit increases timeout and memory risk and makes stage-level monitoring difficult.
A large retry unit increases timeout and memory risk and makes stage-level monitoring difficult.
State transitions express dependencies and provide controlled Retry and Catch behavior.
State transitions express dependencies and provide controlled Retry and Catch behavior.
Lifecycle manages object transitions and expiration, not ETL dependencies.
Lifecycle manages object transitions and expiration, not ETL dependencies.
Dependencies and failure ordering are distributed and duplicate runs are difficult to control.
Dependencies and failure ordering are distributed and duplicate runs are difficult to control.
Check official documentation
Compare the following points with the official documentation.
AWS Step Functions Amazon States Language、Task、Retry、Catchの公式説明を確認する。
What to confirm
時刻起動の寄せ集めと、依存関係を持つワークフローの違いを説明できる。
Key points
- Step Functions
- Dependency
- Retry/Catch
- Idempotency
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
Failure boundary
Retry temporary faults but route permanent input failures to an explicit path.
Safe replay
Each stage should tolerate the same input being delivered more than once.