Metrics Dashboard

The Metrics page (/metrics) shows aggregated performance statistics computed directly from results ingested into QA Hub — no external TestOps provider required.

How metrics are computed

Metrics are derived from NativeTestResult rows — raw results ingested via /api/v1/ingest or the CLI. They are aggregated on demand, not pre-computed, so they always reflect the most recent data.

Available metrics

Pass rate

Overall percentage of PASSED results across all ingested runs in the selected period. Breakdowns are available per project, framework, and environment.

Flakiness ratio

The proportion of test cases that have alternated between PASSED and FAILED in recent runs. A test is considered flaky if it has both a pass and a fail within the window.

Duration percentiles (P50 / P95)

Execution time distribution across all results with a recorded duration_ms:

• P50 — Median test duration; baseline for "typical" test time
• P95 — 95th percentile; reveals slow outliers that inflate CI time

By framework

When results include a framework label (e.g., playwright, cypress, jest), metrics are broken down per framework so you can compare stability across your test stack.

Filters

Filter	Options
Period	Last 7 days, 30 days, 90 days
Project	All projects or a specific one
Framework	All frameworks or a specific one
Environment	All environments or a specific label

Data retention

Native test results are retained for a configurable number of days (default: 30). Older results are pruned by the /api/v1/admin/sweep endpoint, which can be called on a schedule from your CI/CD pipeline.

# Trigger a sweep via the REST API
curl -X POST https://your-qahub.com/api/v1/admin/sweep \
  -H "Authorization: Bearer qh_xxxx_yyyy"

Difference from Coverage Analytics

Feature	Metrics Dashboard	Coverage Analytics
Data source	Native ingestion only	Native or TestOps
TestOps sync required	No	Optional
P50/P95 duration	Yes	No
Environment health bands	No	Yes
Executive health summary	No	Yes