Count clicks per user in five-minute windows.
Keep prior events while the window is open and emit the aggregate when the window closes.
Which processing approach is most appropriate?
Use stateful stream processing with a keyed time window.
Detailed explanation
A stateless function does not retain prior events needed for a five-minute per-user count.
A stateless function does not retain prior events needed for a five-minute per-user count.
Repeated full-history reads add latency and cost and are not low-latency stream processing.
Repeated full-history reads add latency and cost and are not low-latency stream processing.
The processor retains keyed intermediate state and emits the aggregate at the window boundary.
The processor retains keyed intermediate state and emits the aggregate at the window boundary.
CloudTrail audits AWS API calls and is not an application-event windowing engine.
CloudTrail audits AWS API calls and is not an application-event windowing engine.
Check official documentation
Compare the following points with the official documentation.
AWSのストリーム処理における時間窓、状態管理、再処理の公式ドキュメントを確認する。
What to confirm
ステートレス変換と、時間窓・キー単位の状態を持つ処理の違いを説明できる。
Key points
- Stateful processing
- Time window
- Intermediate state
- Stream aggregation
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
Stateless versus stateful
Record-by-record transforms do not retain history; windows, deduplication, and session logic require state.
Late and duplicate events
Windowing designs should define handling for late arrivals, state recovery, and repeated delivery.