Database Migration Guides

Modern database migration isn’t a single practice — it’s a discipline spanning safety, collaboration, and scale. The decision that shapes everything else isn’t which tool you choose, but how you model the problem: do you declare what the database should look like and let the tool compute the change (state-based / declarative), or hand-write every alteration as an ordered script (migration-based / imperative)? That one choice reframes how drift happens, how rollback works, and how teams collaborate. The guides below are organized around three of those concerns:

• Reliability & safety — Drift, failed-migration recovery, and zero-downtime deployment all rest on predictability: a declared target means every environment converges to the same state, drift surfaces instead of hiding, and rollback is redeploying a known-good package — no migration ledger to unwind.
• Workflow & collaboration — Putting schema definitions in source control (database as code) means reviewers see the table as it will be, not a string of mutations — and DevOps integration and approval workflows add the automation, audit trails, and separation of duties that regulated teams need.
• Architecture & sequencing — Across dependencies and multiple environments, state-based deployment resolves much of the ordering for you — foreign keys applied after their tables, interdependent objects retried until they settle — so one package lands the same way in dev, staging, and production. Genuinely circular dependencies are the case you design around, and the dependency guide shows how.

Start anywhere below — each guide goes deep on its corner of the discipline.