Quick Reference
Open the dashboard
- Open your organization, then select the project
- Pick a time range from the selector in the top right: Last 7 days, Last 30 days, or Last 90 days
- Use Refresh to pull the latest numbers after a run finishes
A new project shows No Test Runs Yet until its first run finishes processing. Upload a run, then refresh.
Metric cards
4 cards across the top, all scoped to the selected time range. Every card has an info icon with the same definition you see here.
Run Pass Rate is strict on purpose. It answers “can I trust a green run” rather than “how many tests passed”. For per-test pass counts, read the Pass / Fail Trend chart instead.
Key takeaway: Reliability Score is the number to watch week over week. Run Pass Rate moves on a single bad run; the reliability score weighs every test’s behavior across the range.
Branch Summary
Every branch with runs in the selected range, busiest first. Each row shows the branch name, its environment badge when branch-to-environment mapping is set up, how many runs it had, how long ago the last one was, its pass rate, and the status of that last run.
The pass rate here uses the same strict rule as the Run Pass Rate card, scoped to that branch. Use it to spot a feature branch that is failing while main stays green, or the reverse.
Branches are not capped. If the list is long, scroll inside the card.
Most Flaky Tests
Tests that pass on one run and fail on another, ranked by how often it happens. The card header counts how many tests are flaky and how many flaky incidents happened in the range.
Each row shows its rank, the test case name, its spec file, the branch and environment where it last went flaky, and a flaky rate badge. The badge color tells you how bad it is:
Click a test name to open that test case at its most recent flaky run, where the trace, error, and retry history are. Use the search box to filter by test name. The top 10 are listed.
Learn how detection works and how to fix the common causes in Flaky Test Detection.
Degrading Tests
Tests whose duration is trending upward, ranked by how fast they are slowing down. The badge shows the rate of increase, in milliseconds or seconds gained per hour of elapsed time.
Only passing runs count toward the trend, so a test does not land here because it started failing and timing out. It lands here because the same successful test keeps taking longer. That usually points at a growing fixture, a slower page under test, or a wait that no longer fits.
Small wobbles are filtered out. A test appears only when it has enough runs to establish a trend and the drift is large enough to matter against its own average duration, so a 20 ms creep on a 30-second test does not crowd out a real regression.
Click a test name to open its latest run. Search filters by name. The top 10 are listed.
Chronic Failures
Tests that keep failing without recovering. These are different from flaky tests: nothing about them is intermittent, they are red and staying red.
Each row shows the test, its spec file, the branch and environment of its last failure, and a run streak badge: the number of runs it has failed back to back. A test is flagged as chronic when it has failed 5 or more runs in a row, when it fails at least 70% of its runs, or when a failure streak has been running for 2 or more weeks.
Skipped runs do not break a streak. A run where the test passed does.
Click a test name to open its most recent failure. The top 10 are listed, longest streak first.
Recent Fixes
Tests that were chronic or flaky and have since passed cleanly. Each row shows the test name and how long ago it recovered.
A test is counted once, at the run where it recovered, so the card reads as a list of wins in the range rather than a running total. It is the counterweight to Chronic Failures: after a fix lands, confirm the test shows up here instead of re-checking it by hand. The 10 most recent are listed.
Pass and Fail Trend
Passed and failed test counts across the most recent runs, oldest on the left. The x-axis is run number, not date, so a busy day and a quiet day take the same width.
Hover any point for the passed count, failed count, and total for that run. A rising failed band across several runs points at a change that broke a group of tests. A rising total with a flat failed band means the suite grew.
This chart covers the last 30 runs regardless of the range you picked at the top, so it stays readable on a project that runs tests hourly.
Test Runs
Full run list with filters and detail tabs
Test Analytics
Custom dashboards for reliability and performance
Flaky Test Detection
How flaky tests are detected and fixed
Test Case History
Per-test pass, fail, and duration history