Deploying a workflow version to a runner

In CloudRadial AutomationAI, a deployment binds a published workflow version to a runner and an environment so the workflow can execute. This article explains what a deployment is, how to create one from the Deployments page, and the Run and Retire actions. It is for the Admin and Owner roles.

We'll follow four main steps:

1. Understanding what a deployment is
2. Creating a deployment
3. Running a deployment
4. Retiring a deployment

Step 1: Understanding What a Deployment Is

Why are we doing this? A deployment is the thing that actually runs, so knowing what it binds together explains the fields you fill in next.

A deployment pins three things together so AutomationAI knows what to run and where:

• A published workflow version, the immutable snapshot that executes (drafts and test versions cannot be deployed)
• A runner, the execution agent in your Azure subscription that picks up and runs the work
• An environment, the named target the deployment runs against

A workflow has one active deployment at a time. To change the version or runner a workflow uses, edit its existing deployment rather than creating a second one.

Step 2: Creating a Deployment

Why are we doing this? This is the step that makes a published workflow runnable on your runner.

Open Deployments and click Deploy, then set:

• Workflow, the workflow to deploy (only workflows without an active deployment appear in the list)
• Published version, defaults to the latest published release version; a workflow with no published release version shows a hint to publish it in the Designer first
• Runner, an active runner to bind to; if you have none, register one on the Runners page first
• Environment, the target label (defaults to prod)

Click Deploy to create it. The new deployment appears in the grid with an active status.

Step 3: Running a Deployment

Why are we doing this? This is how you actually launch a run of the deployed workflow.

The Run action launches a run against the deployment's pinned published version. It is available from the Designer's Run action, where you choose the target deployment to launch. A run does not execute on the control plane: AutomationAI queues the work, and the next time the deployment's runner polls in, it picks up the work and begins executing. Because of this, a run starts as pending and the deployment's status and last-run time update once the runner reports back.

Review the result in History. See Running a deployment and reading run history.

Step 4: Retiring a Deployment

Why are we doing this? Retiring stops a deployment from taking new runs, for example when you are replacing or decommissioning it.

The Retire action stops a deployment accepting new runs. Open the deployment from the Deployments grid and choose Retire.

Important -> Retiring is permanent and cannot be undone. If the deployment is the workflow's last active deployment, retiring it also permanently deletes that workflow's scheduled tasks, which removes its Routine. AutomationAI warns you and shows how many scheduled tasks would be deleted before you confirm.

If you are still having trouble, we're here to help! Submit a ticket here for assistance, and don't forget to check our status page to ensure there are no outages in your area.