Databricks details how Snorkel uses Lakebase branching for per-PR test databases
AISnorkel AI used Lakebase branching to give each pull request its own copy of its roughly 1 TB production Postgres database, cutting provisioning from about 20 minutes to under 60 seconds at p95. The setup keeps the main branch synced with production via AWS DMS, creates a TTL-bound branch per PR, and authenticates CI with service principals and short-lived OAuth tokens.
AIWhy it matters.
The case shows how copy-on-write branching and TTLs can replace slow database restores in CI, with concrete provisioning, storage, and security design details.
Source details
Databricks Blog · databricks.comPublished · added here
Original title: Lakebase branching for CI/CD: lessons from 100 PR databases