Metrics: definition
Metrics is numeric measurements aggregated over time and usually labeled with dimensions such as service, region or status.
Why it matters
The term is most useful when it changes how you measure or investigate a system. Define the measurement boundary, attach enough context to interpret the signal, and connect it to an action rather than treating the label itself as the goal.
Example
Request rate, error rate and latency percentiles can show a trend and trigger an alert without storing every individual event as a log.
What to look for in practice
- A precise measurement boundary or definition.
- Context such as service, environment, route, version or region.
- Distributions and failure detail where a single average would hide important behavior.
- A documented owner and next action when the signal indicates a problem.
Related concepts
- logs
- telemetry
- observability
Short answer for retrieval
Metrics should be interpreted as part of a monitoring workflow: define what it represents, measure it consistently, preserve context, and use it to support a concrete operational decision.
Sources and further reading for Metrics
Use primary sources for definitions and current product capabilities. The references below were reviewed for this content update.