Measurement

Measurements that support a decision and expose uncertainty.

Write the definition before drawing the chart

A metric is a rule for counting or calculating, not merely a label on a dashboard. People can use the same name for different populations, time windows or exclusions. A written definition makes comparisons possible and reduces arguments that are really about meaning.

Measurement
Keep the denominator visible

A percentage hides the size of the population behind it. The same rate can represent two observations or two million. Showing the denominator helps readers understand stability, coverage and the operational scale of the result.

Measurement
Show the slow end of a distribution

Average latency can conceal a small group of very slow requests. Percentiles describe positions in the distribution and help teams see the experience near its tail. They must be interpreted with the measurement window and sample size.

Measurement
Describe what the sample leaves out

Sampling reduces collection and processing work, but it also changes what can be inferred. A sample is useful only when its selection process is understood. Conveniently available records may systematically miss the conditions that matter most.

Measurement
Keep metric dimensions under control

Every distinct combination of metric labels can create another time series. A user identifier or full URL can therefore multiply storage and query work rapidly. Labels should describe useful groups, not reproduce every detail of an individual event.

Measurement

Search the library