Almost all modern tech stack components play along very well with cloud native workflows. Databases however, still feel like they get a special exception – or “Extrawurst” – excluded from the normal DevOps operations and patterns like pull requests, branching, merging, canary deployments and fast deployment cycles.Properly merging and reverting database schema changes still feels like we were in CVS or Subversion times – especially if many teams and tons of data get involved. We will tell how those DB challenges got addressed at GitHub and how the same tooling found its way into PlanetScale’s database branching and merging workflow. We will do a live demo showing how to create free databases on our SaaS platform, branch off from the DB main branch, connect to the DB feature branch from various frameworks and programming languages and propose MySQL schema changes in the associated deploy requests. We will also demo how to test, approve, merge and revert database branches directly from a GitHub pull request along with the corresponding application code changes.

Finally, you will learn about the Open Source technologies that make all of this possible: Vitess and gh-ost, used in production at places like GitHub, MyFitnessPal and Twitter.