Prompt
metrics storage retention
Latest observation
Oct 16, 2026
OpenAI APIWeb search: off
Do you mean how long to keep metrics data in a storage system, or are you asking about configuring retention for a specific metrics backend?
If you want a general definition:
- Metrics storage retention = the period metrics are kept before being deleted or downsampled.
- Common retention windows:
- Short-term: 7–30 days for high-resolution data
- Medium-term: 3–12 months for aggregated data
- Long-term: 1+ years for compliance/trending, often with downsampling
- Retention is usually controlled by:
- Time-based deletion: remove data older than X days
- Tiered storage: keep recent data in fast storage, older data in cheaper storage
- Downsampling/aggregation: reduce granularity over time
If you meant a specific system, tell me which one, for example:
- Prometheus
- Grafana Mimir / Cortex
- InfluxDB
- Datadog
- CloudWatch
- Elasticsearch/OpenSearch
- VictoriaMetrics
and I can give the exact retention settings or best practices.