Skip to Content

Analytics

Analytics answers the question that raw CI logs cannot: is this API behavior getting better, worse, or noisier over time?

In the current webapp, analytics is not one single top-level module. The signal is split across the Home dashboard, Target tabs, Reports, and Performance:

SurfaceWhat it reads
Home DashboardRecent functional run pass rate, active alerts, and recent org activity.
Target Overview / Runs / Tests / EndpointsTarget-local health, failures, test history, and endpoint behavior.
Target MetricsCustom functional metrics emitted by ctx.metric().
ReportsGenerated health digests for an org or project scope.
PerformanceLoad-run latency, throughput, error rate, thresholds, and capacity for one Target.

The Signals Most Teams Use

  • Pass rate trend to spot regressions
  • Flaky tests to find unstable tests
  • Top failing tests to prioritize cleanup work
  • Failure distribution to separate assertion bugs from timeout or network noise
  • Endpoint latency to find performance regressions from trace data
  • Test duration trend to catch tests that are getting slower

These are functional-suite trends, built from uploaded test and contract runs. They are distinct from the load-test Performance surface: a test duration trend tracks how long a test takes; load-test latency, throughput, and capacity live under Performance.

Custom metrics

If your tests emit custom metrics through ctx.metric(), Cloud shows them on the Target’s Metrics tab.

Good uses of analytics

  • deciding which failures matter most
  • proving a flaky suite is actually improving
  • detecting a slow endpoint before it becomes a production issue

Next

Last updated on