Process Hub · Publish, deploy and test
How to publish and deploy a process
Release structural definitions to the runtime after checking dependencies.
On this page
Before you start
- Have a reviewed saved draft, Publish solutions authority and a tested profile runtime SQL target. Coordinate changes with process users and automation/integration owners.
Follow these steps
Screenshots use fictional training data. Names, dates and amounts are examples. Select an image to view it at full size.
- Open Settings → Process Hub → Builder and the intended draft. Check profile, process name and version.
- Open Validate & publish. Review the runtime preview and Publish checks, then correct every blocking issue and save the draft.

Review every publication check and fix validation errors before publishing. Deployment is a separate step. - Publish the reviewed structural version using the offered publish action. Note the published version identifier.
- Open Settings → Process Hub → Deployments. Compare Current version and Deployed version for the process.

Compare published and deployed versions. Review the generated plan before deploying to the approved runtime target. - Select Plan to inspect schema/storage changes and blockers in Latest plan. Review active automation impacts with its owner; resolve them rather than forcing an incompatible structure.
- Select Deploy when the plan is ready and permitted. Wait for success and check that the Deployed version/target status matches the intended version.
- Test the process through the member runtime, then review process grants. Publish automation changes separately when needed; structural deployment does not automatically release workflow drafts.
What you should see
The reviewed published snapshot is deployed to the runtime target and can be used by authorised members.
If something goes wrong
Published but missing from member Process Hub
Check successful deployment, active state, process/dependency grants and module/role access.
The plan reports an automation blocker
Coordinate a compatible automation release or the structural change with its owner before deployment.
You see Clean
Clean is a separate irreversible cleanup operation that can remove runtime data/history. It is not a normal deployment step.