The new tool integrates build management, test and production environments, update condition monitoring, backup processes, and operation history within a single software loop.
Before updating the production environment, the system automatically verifies eight specific conditions, including build origin, test deployment status, testing time, and the availability of a rollback point. If any condition is violated, the update is blocked unless explicitly confirmed, with such actions logged and sent to a corporate bot.
The orchestrator manages database structure changes separately. If a schema change occurs after an update, the system restores both the application and the database during a rollback. Backups are created before irreversible operations, and rollback points are stored directly on the environment to ensure recovery even if the orchestrator is unavailable.
The system supports two delivery types: Java-based WMS applications and Yard container images, with builds created from Git. Each version tracks the commit ID and a list of changes. Deployment is performed without requiring manual engineer connection to each server, followed by a service state check.
Test environment updates can be run automatically, while production updates are scheduled, for instance, during the night. All operations, including builds and backups, are recorded in a unified log and managed via a web interface, API, and a corporate bot.
