Runner Draining & API Indexing
This release improves infrastructure lifecycle management with automatic runner draining. Runners in draining mode now gracefully migrate sandbox backups to a new runner before decommissioning, ensuring zero data loss during scale-down or maintenance operations. Errored sandboxes are archived through the backup flow automatically.
API performance receives several indexing improvements, including indexes on sandbox auth tokens and usage period lookups. Build info querying is also optimized, and public and internal health checks are now separated for more granular monitoring.
On the dashboard side, the snapshot creation form has been redesigned with improved validation and a cleaner dialog structure. The CLI also gains concurrent sandbox deletion support for faster bulk operations.
Release Details
Features:
api: index on sandbox auth token (#3711)
api: automatic runner draining (#3699)
api: optimize build info querying (#3714)
api: usage period sandbox id endAt index (#3712)
api: separate public and internal api health checks (#3725)
cli: sandbox deletion concurrency (#3721)
dashboard: update create snapshot form, dialog & validation (#3691)
Fixes:
api: 500 on sandbox delete (#3716)
api: reduce job log severity to debug (#3720)
api: filter out snapshot regions that are not available to org (#3710)
Refactor:
api: pre-deploy and post-deploy db migrations (#3661)
Chores: