Change control is a feature that manages the configuration of control rules and their scopes during various stages of change, such as submission, publishing, and maintenance. It conducts validations prior to implementing changes, determining whether the stage's constraints are met based on the results. This process is essential for preventing unintended data modifications or deletions that could disrupt downstream business processes. Change control consists of two main components: change rules and change policies:
-
Change rules are composed of metadata that abstract the fundamental properties of objects, which are then paired with either system-generated rules or custom code logic conditions. Objects that do not pass these validation rules are controlled according to the associated change policies.
-
Change policies define the scope and methods of control based on the established change rules. A single policy can encompass multiple rules, each with its own designated control method.
Dataphin currently offers a publishing control feature, which performs checks against user-defined rules and policies before the publishing of tasks to ensure compliance with publishing criteria. Tasks that do not pass these checks are either prevented from being published (strong control) or allowed to publish with a warning (weak control), thereby contributing to the platform's stability.
Publishing control is commonly applied in scenarios including, but not limited to:
-
Time node control: Prevents the publishing of new tasks or updates to existing tasks during critical time periods, such as fiscal year-end, financial audits, major sales events like Double 11, or during cluster transitions, to avoid data calculation discrepancies.
-
Development specification control: Enforces naming conventions for tasks, such as requiring tasks within the ODS project to start with the prefix
ods_in order to be published. -
Account control: Limits publishing permissions to specific users, for example, barring tasks submitted by former employees from being published.
Upon completion of rule configuration, administrators can utilize flexible switches to manage the effectiveness of rules and policies, allowing for tailored control impacts. This capability reduces the need for manual oversight and further solidifies the standardization and stability of platform operations.