CI/CD
Trigger Drizz test plans from GitHub Actions, Jenkins, GitLab CI and others — and what a pipeline cannot do yet.
Any CI/CD system that can make an authenticated HTTP request can start a Drizz run: authenticate, upload the build, trigger the plan.
Platforms
GitHub Actions · Jenkins · GitLab CI · Bitbucket Pipelines · Azure DevOps
Requirement
Anything that can make an authenticated HTTP request
Env vars
DRIZZ_BASE_URL, DRIZZ_API_KEY
Pipeline does
authenticate → upload → trigger
Watch out
The trigger step succeeds even if every test later fails
Prerequisites
Client ID, client secret, auth host and audience, issued by Drizz
API base URL for your organization
Test plan ID, from the web app
A CI secret store holding all five values
curlandjqavailable on the build agent
What a pipeline can and cannot do
The trigger step goes green as soon as the run is triggered
It reports the trigger, not the test outcome
A build cannot be failed on a Drizz test failure
No run status is available to the pipeline
Results are read by a person
Open the report in the Drizz web app
Copy this
Pipeline steps
Request an access token.
Upload the build — only when the app version changed.
Trigger one test plan, or a batch.
Record the
execution_idin the build log.Archive the trigger response as a build artifact.
Open the report in the Drizz web app to read the outcome. This step is manual.
Supported platforms
The integration is identical on every platform. Only the surrounding syntax changes.
GitHub Actions
Pull requests, merges, manual dispatch
Jenkins
Scripted or declarative pipelines
GitLab CI
Branch pipelines, scheduled jobs
Bitbucket Pipelines
Build or release stages
Azure DevOps
Release pipelines
Platform examples
Both examples trigger a run and finish. Each passes whether the tests pass or fail. The step that would wait for the result is marked in the file and deliberately not written, because there is no endpoint to call.
Secrets
Client ID, client secret, audience, auth host, base URL
CI secret store. Never committed
Access token
Run-scoped variable, masked in logs — ::add-mask:: in GitHub Actions
Common mistakes
Treated a green trigger step as a green test run
It only means the run started
Put the Drizz step in front of a release gate
The gate lets everything through
Uploaded the build on every commit
Slow pipelines, and 409s on unchanged versions
Cached the token between pipeline runs
It expires after 24 hours
Echoed the token in a build log
Anyone with log access can trigger your runs
Looped single triggers over 30 plans
HTTP 429. Use the batch endpoint
Next
API overview — the base URL and the auth model
Trigger a run — single and batch triggers
Errors & limits — rate limits and retries
Last updated: 6 August 2026
Last updated
Was this helpful?