tiered retention
Keeping recent data in fast storage and older data in cheaper, slower, often downsampled storage.
In more detail
Each tier should carry a query promise and a recovery promise, and deletion is part of the capacity model. Older tiers are often downsampled and stored in cheaper object storage.
Related terms
- retention — How long data such as backups, logs or snapshots is kept before it is deleted.
- downsampling — Replacing detailed old data with coarser summaries, such as one value per hour instead of per ten seconds.
- archive — Long-term storage for old data that is rarely read, trading query speed for low cost.
Taught in
- Worked Design Three: Scaling The Metrics Pipeline — Scalable System Design & Distributed Architecture