When a stage’s pull request is merged, the new code still has to reach your servers before the run should move on. A stage’s deploy confirmation setting makes the stage wait for exactly that: it holds after the merge until the deployment is confirmed. This page covers turning the setting on or off, what the wait looks like, and how to push past it if you need to.
For what stages and gates are, see The pipeline: stages and gates.
Turn deploy confirmation on or off
Deploy confirmation is set per stage, on the project’s Pipeline page.
- From your project, select Pipeline.
- Select the stage you want to change. Its Stage Configuration panel opens.
- Find Wait for deploy confirmation and switch the checkbox on or off.
- Select Save, and confirm in the Save Pipeline dialog.
A stage you add by hand starts with this setting on. Stages loaded from a pipeline template start with it off.
Turn it on for a stage that deploys to a real environment — production, staging — where the run shouldn’t advance until the deploy has landed. Turn it off for a stage with nothing to deploy, such as a library or a development-only stage.
What the wait looks like
When deploy confirmation is on and you merge the stage’s pull request, the stage doesn’t pass straight away. Instead:
- The stage enters the Waiting for Deploy state — an indigo badge on the run detail page and the task page. See Stage states.
- The run as a whole stays in the Waiting for Gate run state.
- The stage holds there until DIJJI.ai is told the deployment has finished.
How the wait clears
The wait clears when DIJJI.ai receives a deployment confirmation for the merged code. That confirmation comes from your deployment setup — once the deploy of that change completes, DIJJI.ai picks it up and the stage moves on to Passed, exactly as a merged-and-deployed stage should.
There’s a short delay between the deploy finishing and the run catching up. Give it a moment and refresh the run detail page if it still shows the old state.
Force Advance — when the confirmation never comes
Sometimes a deployment confirmation never arrives — the deploy isn’t set up to report back, or something went wrong on the deployment side. So a stage doesn’t wait forever, DIJJI.ai gives you an escape hatch.
One hour after the stage’s pull request was merged, a Force Advance button appears on the task page, on the card for the waiting stage. Selecting it treats the deployment as done and moves the run on to the next stage.
Use Force Advance when you’ve confirmed the deploy yourself, or decided the run should proceed regardless. It’s a manual override, not a retry: it doesn’t check anything, it simply releases the stage. Force Advance isn’t admin-only — anyone on the project’s team who can open the task can use it.
Force Advance is the escape for a deploy wait. A stage held on its E2E tests has a different one — a Skip E2E button on the stage card, which marks the tests skipped and advances the run. See Trigger E2E tests manually.
What’s next
A stage past its deploy wait carries on through the rest of the pipeline. For the states a run and its stages pass through, see Run states and Stage states.