GitLab CI/CD Setup

You'll need

• a GitLab project with a test runner reporter already configured (see Getting Started)
• the project is built on the GitLab instance configured for your Flakiness.io deployment
• your Flakiness.io project linked to that GitLab project

You'll have

• CI uploading reports to Flakiness.io on every push and merge request, using a short-lived token exchanged via GitLab OIDC.

On GitLab CI/CD, reporters can authenticate via GitLab OIDC, so there are no secrets to create, rotate, or leak.

Setup

1. Confirm the reporter has flakinessProject set. OIDC requires the project identifier so Flakiness.io knows where to route the upload. If you followed Getting Started, this is already done.

   What is flakinessProject?

   The orgSlug/projectSlug identifier from your project’s URL. For example, a project at https://flakiness.io/my-org/my-app has flakinessProject: 'my-org/my-app'.

2. Declare an ID token named FLAKINESS_ID_TOKEN. Its aud must be your Flakiness.io project identifier.

   .gitlab-ci.yml

   test:
     stage: test
     id_tokens:
       FLAKINESS_ID_TOKEN:
         aud: my-org/my-app
     script:
       - npx playwright test

3. Push. The reporter picks up the ID token automatically at upload time.

On success you’ll see a confirmation line in the test output:

[flakiness.io] ✓ Uploaded as https://flakiness.io/my-org/my-app/run/42

Self-managed GitLab

Hosted Flakiness.io verifies ID tokens from https://gitlab.com. A self-hosted Flakiness.io deployment can instead set GITLAB_ORIGIN to its self-managed GitLab instance. The same OIDC configuration then works because Flakiness.io uses that instance as the token issuer and reads its signing keys. The GitLab origin must be reachable from the Flakiness.io container.