The model provider announces a new version.
Quality and safety may change after the update.
Choose two appropriate change-management practices.
Compare old and new versions with a fixed set and define a rollback condition and procedure before switching.
Detailed explanation
The application can return safely if the update causes problems.
Limits and output tendencies may change.
Quality, format, safety, cost, and latency differences become visible.
Root-cause analysis and reproducibility are lost.
Check official documentation
Compare the following points with the official documentation.
Amazon Bedrock公式モデルライフサイクルとAIF-C01 Domain 3.1・5.2を確認する。
Official references
What to confirm
モデル更新を通常の変更管理として扱う理由を説明できる。
Key points
- Versioning
- Evaluation
- Rollback
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Model versioning
Tracking an immutable model artifact, its settings, evaluation evidence, and release history.
Canary release
Exposing a new version to a limited share of traffic before broader rollout.