Storage
Amazon S3 Lifecycle policies
Amazon S3 Lifecycle policies transition or expire objects according to age, version state, tags, and other supported filters. They enforce retention economics for prompts, logs, corpora, and evaluation artifacts without periodic cleanup code.
Key points
- Rules can target current versions, noncurrent versions, and incomplete multipart uploads
- Minimum storage duration charges apply to several archival classes
- Expiration markers and versioning behavior must be designed together
When to use it
- Archive old evaluation datasets after an audit window
- Expire raw model interaction logs after the approved retention period
Exam tips
- Choose Lifecycle policies for predictable age-based handling and Intelligent-Tiering for uncertain access
- Validate legal holds and version retention before enabling expiration