Watch a Build
3 min read
Opening an automated test run shows the automation view rather than the manual one — there is nothing to record by hand, because the pipeline records it.
The page has three parts: the automation panel at the top, a summary underneath it, and the build history below that.
The automation panel
Section titled “The automation panel”The panel names the run, the provider driving it, and the pipeline it is bound to — repository and workflow, with the branch alongside. Use Select to attach a pipeline, or Change to point the run somewhere else.
Two actions:
- Run starts a new build on the provider.
- Import existing build opens “Import results from an existing build” so you can pull the report from a build that has already finished.
Before a pipeline is attached, the panel prompts: “Choose the repository and workflow that runs this test run’s cases, then press Run.”
Two states disable the actions:
- The run is completed or its release is locked — “This run no longer accepts results, so automation is read-only. Resume the run to import again.”
- You lack permission — “You don’t have permission to start or import builds on this project.”
While a build is in flight the page refreshes on its own every few seconds, fetching only what changed.
Run states
Section titled “Run states”Hawzu tracks the build’s state on the provider separately from the run’s state — the work Hawzu itself is doing. That separation is deliberate: “your tests failed” and “we could not import your results” are different problems.
| State | Meaning |
|---|---|
Queued | Hawzu has asked the provider to start a build |
Build running | The build is executing on the provider |
Downloading report | The build finished; Hawzu is fetching its artifacts |
Parsing report | Hawzu is reading the results out of the report |
Importing results | Results are being mapped onto test cases |
Completed | Results are in the run |
Failed | Hawzu could not get results — the reason is shown on the build |
Cancelled | The build was cancelled |
Hawzu polls the provider on a backoff — 5, 10, 20, 30, then 60 seconds — so a long build is not hammered, and a short one is picked up quickly. A build that has not finished after 12 hours is abandoned with “Gave up waiting for this build to finish.”
Summary
Section titled “Summary”Four tiles sit above the history: Latest build, Latest build results (pass rate with a status meter), All build results, and Builds.
Build history
Section titled “Build history”The Build history section lists every build for the run, newest first, with columns for Build, Pipeline, State, Results, Duration, and Notes. Expand a row — several at once, if you are comparing — to see:
Pipeline config — the repository, workflow, branch, and build parameters that build used. Useful when a run’s pipeline has been changed since.
Source report — download links to the raw report artifacts Hawzu archived for that build, so you can check what Hawzu actually read. If archiving failed, the panel says so and gives the reason.
Failure clusters — failures grouped by signature, shown as “N failures, M causes”, each expandable to the tests inside it. Twenty failing tests with one root cause is a much shorter read than twenty stack traces. The section is hidden when grouping would not help.
Results — every row the report produced, with Test, Status, Outcome, and Duration, and expandable stack traces. Applied rows come first.
| Outcome | Meaning |
|---|---|
| Applied | Matched a test case in the run and set its result |
| Not in this run | Matched a test case, but that case is not part of this run |
| No matching test case | Nothing in the project matched this row |
Rows that are not Applied are kept for reference only. They never affect status counts, analytics, test case history, or release readiness. See How Results Map to Test Cases.
Builds that overlap
Section titled “Builds that overlap”If a build finishes after a newer build has already reported, its results are marked superseded: kept for reference, not applied. The build carries the explanation — “A build started after this one had already reported, so this build’s results were kept for reference and not applied to the run.”
A build that matched nothing at all leaves the run untouched. Hawzu will not wipe a run’s results because a report was empty or unreadable.