convex-migration-helper

To add a required field to a Convex table without downtime, use a widen-migrate-narrow workflow that temporarily widens the schema, backfills existing documents with the new field, and then narrows the schema to enforce the constraint safely.

The widen-migrate-narrow workflow in Convex is a multi-deploy strategy for breaking schema changes. It widens schema constraints to accept both old and new shapes, migrates and backfills data, and then narrows constraints to the desired strict state.

You can run batched data backfills in Convex using the @convex-dev/migrations component, which provides resumable, monitored data transformations for live applications without interrupting normal operations.

Yes, @convex-dev/migrations supports zero-downtime rollouts for Convex type changes by implementing dual-write, dual-read, and dry-run verification strategies alongside rollback-aware deployment sequencing.

The best way to handle Convex table splits and merges in production is using a structured migration plan with schema validation awareness, batched resumable backfills, and dry-run verification to prevent data corruption or failed deployments.

You should avoid manual schema updates in Convex when dealing with breaking changes like field deletions or type changes on live applications, as unmonitored updates risk corrupting production data and causing deployment failures without rollback safety.