Skip to main content

Metrics & Events

Unified Monitoring uses a common data model for metrics (numerical time series) and events (discrete occurrences such as alerts, changes, or incidents). This model abstracts away source‑specific details so you can define policies, dashboards, and reports consistently.

This section explains how metrics and events are represented and how to design a label strategy that balances observability with scale.

Metrics vs. Events​

  • Metrics – values sampled over time, such as CPU utilization, request rate, error count, or latency. Metrics are ideal for SLOs, dashboards, and threshold‑based alerting.
  • Events – records of something that happened at a point in time, such as an alert firing, a deployment, a configuration change, or a security finding. Events are useful for correlation and audit trails.

Both metrics and events are linked to CMDB CIs where possible, so that you can move between resource records and their associated telemetry.

Labels, Tags, and Cardinality​

Unified Monitoring relies on labels/tags to provide business and technical context, such as account, region, environment, application, service, and owner. A well‑designed label strategy is critical for:

  • Filtering dashboards and alerts by meaningful attributes.
  • Aggregating metrics by application, environment, or business unit.
  • Avoiding explosive cardinality (too many distinct label combinations) that can impact performance and storage.

See Labels & Tags for recommended patterns.

Retention and Rollups​

Monitoring data is stored at multiple granularities:

  • High‑resolution samples for recent periods (for example, 1–5 minutes).
  • Aggregated rollups (for example, hourly or daily summaries) for longer‑term analysis.

Exact windows depend on your subscription and environment. For conceptual guidance, see Retention & Rollups and Limits & Quotas.

Reading Path​

Metrics​

Understand metric types, naming, dimensions, tags, and CMDB context for time-series monitoring data.

Events​

Learn about event sources, event structure, and how events correlate with metrics during investigation.

Labels & Tags​

Design a label and tag strategy that supports filtering, aggregation, ownership, and scale.

Retention & Rollups​

Understand data tiers, rollup behavior, and the impact of retention choices on dashboards and alerts.