Skip to main content

Retention & Rollups

Unified Monitoring stores monitoring data at different granularities over time to balance troubleshooting needs, trend analysis, and storage cost.

Use this guide to review design considerations for retention and rollups.

Data Tiers​

Monitoring data is typically kept in:

  • High‑resolution tier – short‑term storage of raw or near‑raw samples (for example, 1–5 minute resolution) for detailed troubleshooting.
  • Rollup tier – aggregated metrics (for example, hourly or daily statistics) stored for much longer periods for capacity planning, reporting, and audit.

Events may have their own retention policies distinct from metrics, especially for compliance or security findings.

Design Considerations​

When planning retention:

  • Identify how far back you need to troubleshoot incidents in detail.
  • Decide how long you must keep summary data for capacity, budgeting, or compliance reports.
  • Consider regulatory requirements for specific environments (for example, regulated industries may need multi‑year retention).

Impact on Dashboards and Alerts​

  • Dashboards focusing on recent incidents typically use high‑resolution data.
  • Long‑term trend dashboards rely on rolled‑up metrics and may show approximate values.
  • Alert policies should generally be based on high‑resolution data; however, some slow‑moving SLOs can use rollups.

For conceptual limits and common constraints, see Limits & Quotas.