Events arriving in Kinesis Data Streams must be transformed by Lambda while the processed position advances automatically.
Temporary errors should retry the batch, and downstream writes must remain safe if a record is delivered again.
Which design is most appropriate?
Use a Lambda event source mapping for Kinesis, manage checkpoints through the integration, and make downstream processing idempotent.
Detailed explanation
The mapping polls shards and invokes Lambda with batches, advancing checkpoints according to processing results.
The mapping polls shards and invokes Lambda with batches, advancing checkpoints according to processing results.
Shard leases, checkpoints, retries, and scaling would be custom-managed instead of using the managed integration.
Shard leases, checkpoints, retries, and scaling would be custom-managed instead of using the managed integration.
S3 notifications concern S3 objects and do not read Kinesis Data Streams records.
S3 notifications concern S3 objects and do not read Kinesis Data Streams records.
CloudTrail is an audit record for API activity, not the business event stream in Kinesis.
CloudTrail is an audit record for API activity, not the business event stream in Kinesis.
Check official documentation
Compare the following points with the official documentation.
AWS LambdaのKinesisイベントソースマッピング、バッチ処理、チェックポイント、再試行の公式説明を確認する。
Official references
What to confirm
KinesisからLambdaを呼ぶマネージド連携と、S3通知・監査ログ・自作ポーリングの違いを説明できる。
Key points
- Kinesis Data Streams
- Lambda event source
- Checkpoint
- 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 the relevant DEA-C01 Skill and current AWS service documentation.
Foundation review
Batch invocation
Tune the records per invocation and failure behavior to balance throughput and latency.
Retry safety
Assume an event can be delivered again and protect downstream writes from duplication.