Aggregate Kinesis click events per user in five-minute windows.
Retain events while a window is open and update the result for late arrivals.
Which processing platform is most appropriate?
Use Amazon Managed Service for Apache Flink for stateful stream processing and time windows.
Detailed explanation
S3 Select queries stored objects and does not maintain state for a continuous stream window.
S3 Select queries stored objects and does not maintain state for a continuous stream window.
CloudTrail Lake analyzes audit events and is not a click-stream windowing engine.
CloudTrail Lake analyzes audit events and is not a click-stream windowing engine.
Lifecycle rules transition or delete objects and do not execute stream aggregation.
Lifecycle rules transition or delete objects and do not execute stream aggregation.
Managed Flink supports stateful streaming, time windows, and continuous aggregation.
Managed Flink supports stateful streaming, time windows, and continuous aggregation.
Check official documentation
Compare the following points with the official documentation.
Amazon Managed Service for Apache Flinkの状態管理、時間窓、チェックポイントの公式説明を確認する。
What to confirm
バッチクエリと状態付きストリーム集計の違いを説明できる。
Key points
- Apache Flink
- Time window
- State management
- Checkpoint
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
Event time
Late events require an explicit policy for watermarks, allowed lateness, and updates.
Recovery semantics
Checkpoints and idempotent output handling limit duplicates after a restart.