Branching databases like code: a CI/CD pattern for Lakebase, in production at Glaspoort
Summary
Glaspoort has implemented a production-ready CI/CD pattern for Lakebase that branches development and acceptance environments directly from production to avoid the "reset-from-parent trap." This architecture enables a per-PR workflow where disposable branches are tested against live app images, treating migrations as the source of truth and utilizing a velocity-first promotion model backed by stack revalidation safeguards.
Summary generated by brickster.ai. For the full article, follow the source link above.
Topics
More from Databricks Blog
Get Started with Genie One: Top AI Cowork Use Cases for Business Users
Genie One enables business users to automate recurring reviews and meeting workflows by pulling governed data into calendar-aware briefs and consistent reports. By packaging these processes into reusable AI skills and agents, teams can
The EU Digital Product Passport: a traceability deadline
The EU Digital Product Passport makes product traceability a mandatory condition for market access starting in February 2027, turning data availability and quality into the primary implementation challenge for global producers. Databricks provides
Building AI Literacy: Frameworks, Tools, and Practices
A three-domain framework encompassing Understand, Evaluate, and Use offers organizations a structured approach to AI literacy, curriculum development, and governance policies. Developing these critical competencies—including identifying model hallucinations and bias—
AI Applications in Finance: A Practical Use Case Guide
Finance teams can establish a low-risk path to enterprise scale by adopting a staged rollout featuring two prioritized pilots, a 90–120 day evaluation window, and rigorous ROI measurement. Successful deployment
How the FDA Built an AI Platform That 85% of Its Staff Now Use Daily
The FDA broke down center silos to launch an enterprise AI platform that reached 85% staff adoption in just two months, cutting regulatory research times from days to three minutes. Powered by a governed data
Permission isn't purpose: Intent-based authorization in Omnigent
Omnigent introduces intent-based authorization for AI agents, binding each session to a declared, human-approved purpose rather than relying solely on identity-based permissions. Because any action outside this