What is Workflow Versioning?

Once a workflow is live and processes are actively running through it, editing its definition in place is dangerous: a process that started under version 1 might reach a step that no longer exists in version 2, or receive data in a shape the new step doesn't expect. Versioning solves this by treating each edit as a new, separately identified revision, letting already-running instances finish on the version they started with while new instances pick up the latest one. This is conceptually similar to how a database migration or an API contract is versioned rather than mutated silently. Versioning also enables safe experimentation: a team can stage a new version, test it against sample data, and only promote it to be the default for new runs once it's verified, with a clear rollback path if the new version misbehaves in production. Audit and compliance requirements often demand this too, since being able to show exactly which workflow logic processed a given customer's data at a specific point in time is a common evidence request.