convex-migration-helper

Online Convex migrations use a widen-migrate-narrow pattern with batch processing to safely backfill data while keeping applications live. This structured approach allows schema changes to deploy without downtime.

Breaking schema changes in Convex require a multi-step pattern: widening the schema first, backfilling and migrating the data, then narrowing the schema to enforce the new types.

You can backfill Convex fields using the @convex-dev/migrations component, which supports batch processing and resume capabilities to reliably populate new schema fields from existing data.

Yes, you can validate Convex schema migrations using a dry run. This capability allows you to test the data backfill and schema switching logic safely before executing on production data.

Splitting tables or renaming fields requires an online-switching migration strategy. You deploy the new schema alongside the old one, backfill the data across them, and safely switch reads and writes before removing the old fields.