# Migrate Supabase database: Durable job state

## Summary

Move PostgreSQL data from Supabase database to Ample, one component at a time. Destination verified on Ample: a jobs table with a state and attempts counter, processing inside a row-locked transaction, and a second run that is a no-op (job=done attempts=1 idempotent=true). Source procedure: Take a logical backup with `pg_dump -Fc` using the direct (non-pooled) connection string. Not migrated automatically: Supabase Auth, Storage, Realtime, Edge Functions and row-level security policies tied to Supabase roles are not migrated by a database move. Cutover: Freeze writes at the source with the user's go-ahead, take the final dump, restore, validate counts and checksums, redeploy the app with the managed DATABASE_URL, keep the Supabase project until confirmed. Rollback: keep the source untouched until you confirm; nothing at the source is changed or deleted by this guide.

## Infrastructure requirements

- [Postgres](/discover/primitive/postgres): verified (Managed PostgreSQL 16 runs in its own microVM and is auto-provisioned when an app needs a database and no DATABASE_URL is supplied.)

## Prerequisites

- Authorized access to the Supabase database source and its export tooling
- An inventory of every component in scope and out of scope
- A validated backup or copy before any cutover
- An Ample account token with servers:write, databases:read

## Exact tested configuration

- template: `node-22`
- runtime: `node`
- size: `s-1vcpu-1gb`
- install: `npm install`
- build: `npm run build --if-present`
- start: `npm run start`

## Steps

1. **Inventory the source.** List what Supabase database provides beyond the component you are moving. Out of scope here: Supabase Auth, Storage, Realtime, Edge Functions and row-level security policies tied to Supabase roles are not migrated by a database move.
2. **Source step 1.** Take a logical backup with `pg_dump -Fc` using the direct (non-pooled) connection string.
3. **Source step 2.** Check extensions and Supabase-managed schemas (`auth`, `storage`, `realtime`) in the dump; only your application schema moves.
4. **Source step 3.** Restore into a scratch managed database first and compare row counts.
5. **Provision and restore.** Deploy the app once so a managed PostgreSQL 16 database exists (or create one with `ample database create --engine postgres`), then restore the dump with pg_restore using its connection string; keep migrations idempotent.
   
   ```
   ample database create --name <db-name> --engine postgres
   ```
6. **Validate before cutover.** Run the app's own checks and, for data, compare counts and checksums; the example checks are the pattern self-tests (/p/durable-job-state).
7. **Cut over.** Freeze writes at the source with the user's go-ahead, take the final dump, restore, validate counts and checksums, redeploy the app with the managed DATABASE_URL, keep the Supabase project until confirmed.
8. **Rollback.** Point DNS or configuration back to the source. The source was never modified; deletion is a separate, user-executed step after validation.

## Tested examples

- **Express pattern fixture (destination)** (tests/deploy-canaries/express-patterns): Verified destination basis: durable job state.

## Success checks

- destination app responds on its public URL (`/` on the live URL, expect ample canary express patterns)
- durable-job-state check from the destination example (`/p/durable-job-state` on the live URL, expect see the pattern fixture checks)
- data or object counts and checksums match the source

## Limitations

- Documentation only: nothing is executed automatically and no execution binding is offered.
- The source-side procedure is documented from Supabase database's standard tooling and was not executed in this catalog's evidence; the destination side was verified with the pattern fixture.
- No full source-product parity is claimed: Supabase Auth, Storage, Realtime, Edge Functions and row-level security policies tied to Supabase roles are not migrated by a database move.
- Verified on the node-22 template at s-1vcpu-1gb; region, compliance and request-duration limits are unknown.
- Managed PostgreSQL 16 only; extensions, connection limits and backup or restore procedures on the managed side are not verified beyond the pattern checks.

## Cost estimate

Estimated 10.00 USD per month (size prices from pricing.toml at build revision a1b8c38919e59cd035ebabaced73cf84ece24371).
- app server x1 `s-1vcpu-1gb`: 5.00 USD
- managed PostgreSQL database x1 `s-1vcpu-1gb`: 5.00 USD

Destination always-on monthly price of the tested sizes; apps auto-pause when idle. Source costs are unknown to Ample.

## Verification evidence

- canary_run on 2026-09-21T02:34:39Z at revision `1d28ae0-dirty (CLI 0.1.21)`: Destination side verified: the Express pattern fixture deployed on Ample (npm install, npm run build --if-present, npm run start on the node-22 template) and its checks passed (a jobs table with a state and attempts counter, processing inside a row-locked transaction, and a second run that is a no-op (job=done attempts=1 idempotent=true)). The source-side export from Supabase database is documented from the vendor's standard tooling and was not executed by this catalog's evidence. (expires 2027-03-20T02:34:39Z)

## Execution binding

No execution binding. This recipe is documentation only; nothing is executed automatically.