Progressive deploys

Learn how to deploy apps to your readers with a staged approach.

Use progressive deployments to roll out your application to readers within a deploy group in stages. Each stage corresponds to a percentage of the deploy group’s readers that’s greater than the previous stage’s percentage, with the final stage always at 100%. You can use one of our predefined plans or create your own custom plan.

To set up a progressive deployment:

Navigate to the desired app details page, click Deploy version, then select the intended app version and deploy group.
Select a custom or predefined plan under Progressive Deployments.
- After creating the deploy plan, you can’t edit the percentage associated with each stage.
Confirm your deploy plan, then click Deploy.
The deployment starts according to the percentage specified for the initial stage.
- If you selected a predefined plan, the deployment starts at 0%. You need to manually advance to the next stage for any readers to receive the new version of the Terminal app.
To advance a deployment, click Update under Deployed Group on the app details page. Confirm the next stage to advance the plan.
To pause a deployment, click Update to open a drawer, then click Pause. This stops the rollout of your application to any additional readers.

Keep the following in mind when using progressive deployments:

Deployments don’t automatically advance. If you never manually advance the deploy plan, the deployment remains at its current percentage.
Readers are randomly selected within a deploy group for inclusion in a stage of the progressive deploy plan. For example, if Version 2.0 of an app is currently deployed to 40% of Deploy Group A, then 40% of readers in Deploy Group A have Version 2.0 of the app and 60% have the version prior to 2.0.
As the deployment advances, the group of readers receiving the new version includes all readers who received it in the previous stage.
You need to select the desired progressive deploy plan each time you create a new deploy plan. Rollout stages don’t persist between old and new deployment plans.