Release control, governed by user-defined rules and policies, conducts checks prior to task deployment to ascertain compliance with release conditions. Tasks that do not pass these checks are either blocked from release (strong control) or issued a warning without blocking (weak control), thereby enhancing platform stability.
Feature overview
Release control is a component of change control, which encompasses change rules and change policies.
Create change rules: Derive abstract metadata from the fundamental properties of an object, and establish validation rules by integrating built-in system rules or by applying logic conditions through custom code. Objects that do not comply with these validation rules will be managed accordingly, with the method of control defined by the associated change policy.
Create Change Policies: Define the scope and enforcement method for user-created change rules. This allows for the configuration of multiple rules within a single change policy, with distinct enforcement methods for each rule.
Scenarios
Release control is typically applied in scenarios such as:
Time Node Control: To mitigate the risk of data calculation changes, the release of new or updated tasks is restricted during critical time nodes, such as fiscal year-end, financial audits, and during significant events like the Double 11 promotion, ensuring cluster switch operations are not affected.
R&D Specification Control: Releases are regulated based on specific task properties. For instance, tasks within the ODS project must begin with the prefix
ods_to be eligible for release.Account Control: Certain submitters or publishers face restrictions. For example, tasks submitted by former employees are barred from release, and individuals with a Dataphin asset administration module personal health score below a set threshold are not permitted to perform release operations.