Not every team wants Hawzu starting their builds. If your CI already runs on its own schedule — nightly regression, a post-merge suite, a pipeline chained to a deploy — it can tell Hawzu it has finished instead, and Hawzu imports the results from there.
The import path is identical either way. The only difference is how the build’s completion is discovered.
How it works
Section titled “How it works”Each automated test run’s pipeline has a signed webhook URL:
Your pipeline posts a small notification payload when it finishes. Hawzu resolves it to the right test run and queues the same import job the poller uses — downloading the artifacts, parsing the report, and mapping results onto test cases exactly as described in How Results Map to Test Cases.
Authenticating the call
Section titled “Authenticating the call”A CI server has no Hawzu session, so the endpoint takes no login. Authorisation rests on two things, and both must hold:
- The configuration id in the URL.
- An HMAC-SHA256 signature over the raw request body, keyed by that pipeline’s webhook secret, sent in the
X-Hawzu-Signatureheader.
The header value is the hex digest prefixed with the algorithm:
Signatures are compared in constant time, so a wrong one cannot be recovered from response timing.
Request bodies are capped at 1 MB — notification payloads are a few kilobytes, and the cap bounds what an unauthenticated caller can make Hawzu buffer before the signature is checked.
When to use which
Section titled “When to use which”| Situation | Use |
|---|---|
| QA decides when the suite runs | Run from the test run — see Create an Automated Test Run |
| The pipeline already runs on a schedule you do not control | This webhook |
| A provider Hawzu has no adapter for | Importing CI Results — post the report itself |