Have you ever had a database migration fail on production because of a random NULL value somewhere? Ever had a hard time identifying all the possible scenarios in a database migration? Have you ever had to debug a database issue, without having access to the production database? Ever had to wait for a database copy of multiple terabytes worth of relation data? Validating that your database migrations work before running them on production databases should be a standard practice. However, writing test cases that cover all (or even most) cases isn’t easy nor feasible for a multitude of reasons. Data quality in production might not be 100% or not all data might have the constraints that are currently present due to previous migrations. Testing the migration on actual production (like) data is often a cumbersome and slow process, if it’s possible at all. Postgres.ai’s Database Lab Engine provides a way to quickly and easily create copies of PostgreSQL databases. In this talk we’ll show how this can be used to make sure your database migrations don’t fail.