A stream-triggered Lambda transformation is scaling sharply, and downstream RDS connections and API rate limits are being exceeded.
Limit concurrency while shortening processing time and keeping latency observable.
Which Lambda configuration and operating approach is most appropriate?
Set a Reserved Concurrency limit and tune memory (CPU) and event batch size using processing-time, error, and IteratorAge metrics.
Detailed explanation
This can exhaust downstream connections and rate limits and increase latency and duplicate processing.
This can exhaust downstream connections and rate limits and increase latency and duplicate processing.
A function-level concurrency limit controls downstream load, while memory and batch size can be tuned using metrics and processing time.
A function-level concurrency limit controls downstream load, while memory and batch size can be tuned using metrics and processing time.
Lower memory also reduces CPU, which can increase timeouts, backlog, and retries; measure before tuning.
Lower memory also reduces CPU, which can increase timeouts, backlog, and retries; measure before tuning.
Audit-log retention does not control function concurrency or transformation performance.
Audit-log retention does not control function concurrency or transformation performance.
Check official documentation
Compare the following points with the official documentation.
AWS LambdaのReserved Concurrency、メモリとCPU、イベントソースのバッチサイズ、CloudWatchメトリクスの公式説明を確認する。
Official references
What to confirm
同時実行上限、リソース設定、下流のスロットリング、監査ログ保持の役割を区別できる。
Key points
- Reserved Concurrency
- Memory and CPU
- Batch size
- Downstream throttling
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
Resource tuning
Compare memory, CPU, batch size, duration, and errors instead of optimizing one setting in isolation.
Backpressure
Limit or pace consumers so downstream capacity is not overwhelmed.