Performance & Scale

This guide explains how to manage scale as Unified Monitoring coverage grows across accounts, regions, and services so that queries, dashboards, and alerts remain responsive.

Manage Metric and Label Cardinality

• Keep metric names and label sets focused on dimensions you actually query.
• Avoid high‑cardinality labels (for example, user IDs, request IDs, long free‑form strings).
• Use CMDB and log data for highly detailed information instead of encoding it in monitoring labels.

See Labels & Tags for design guidance.

Query and Dashboard Patterns

• Prefer a small number of well‑designed dashboards over many similar variants.
• Use filters and saved searches instead of creating a new dashboard for every application.
• Limit the number of very heavy widgets (for example, many high‑resolution charts over long time ranges) on a single dashboard.

Retention and Storage

• Periodically review which datasets are heavily used vs. rarely queried.
• Monitor storage consumption relative to the limits described in Limits & Quotas.

Operational Monitoring of Monitoring

• Track the health of monitoring pipelines themselves (for example, “no data” alerts for critical sources, ingestion error rates).
• Include Unified Monitoring components in your overall platform health dashboards.