Latency Metrics

Measurements that describe the distribution and extremes of response or processing time.

Motivation

Measurements that describe the distribution and extremes of response or processing time.

Where it fits

This concept provides a shared vocabulary for designing, documenting, measuring, or operating software systems.

Mental model

Treat it as a structured tool: define the question being answered, use consistent terminology, and connect the result to decisions and follow-up work.

Subconcepts

P50 Latency

The median latency: half of observations are faster and half are slower.

P95 Latency

The latency below which 95 percent of observations fall.

P99 Latency

A tail-latency measure below which 99 percent of observations fall.

Mean Latency

The arithmetic average latency across observations.

Max Latency

The largest observed latency in a defined interval.

Practical use

Percentiles are generally more informative than a mean alone because user experience and overload often appear in the tail.

Common mistakes

Using ambiguous definitions or units.
Collecting or documenting information without connecting it to a decision.
Failing to keep the artifact or measurement current as the system changes.

See the linked concepts in the knowledge graph for prerequisites and adjacent practices.