Run db:migrate automatically before every Railway deploy

The 2026-07-06 prod outage was a missed manual migration: the 0002
job_events migration never ran against prod, so the worker crash-looped
on every job. Make the deploy pipeline own it:

- railway.json sets `npm run db:migrate` as the pre-deploy command.
  Railway runs it with the service's env before starting the new
  deployment; a failed migration fails the deploy and the old version
  keeps serving.
- runMigrations() now serializes on a Postgres advisory lock so api and
  worker pre-deploys firing off the same push can't race Drizzle's
  journal writes.
- DEPLOY.md: document the automatic path; keep the manual command for
  first-time setup.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

docs/DEPLOY.md

@@ -12,14 +12,20 @@ Postgres and S3-compatible storage. The recommended OSS-friendly stack:
 ## 1. Database (Neon)
 
 1. Create a Neon project; copy the pooled connection string.
-2. Apply migrations (from CI/CD or locally):
+2. Migrations run **automatically before every Railway deploy**: the repo's
+   `railway.json` sets `npm run db:migrate` as the pre-deploy command, which
+   Railway executes (with the service's env, so `DATABASE_URL` is present)
+   before starting the new deployment. Re-running is a no-op for anything
+   already applied, and concurrent runs (api + worker deploying off the same
+   push) serialize on a Postgres advisory lock. A failed migration fails the
+   deploy — the old version keeps serving.
+
+   To apply migrations manually (first-time setup, or outside Railway):
    ```bash
    DATABASE_URL='postgres://...neon.../ditto_site?sslmode=require' npm run db:migrate
    ```
-   Run this on every deploy that changes `packages/db/migrations`. Drizzle tracks
+   (Alternatively, apply a single migration by piping its SQL through
-   applied migrations in the database, so re-running is a no-op for anything
+   `psql`/the Neon CLI — but prefer `db:migrate`, which also records the
-   already applied. (Alternatively, apply a single migration by piping its SQL
-   through `psql`/the Neon CLI — but prefer `db:migrate`, which also records the
    journal entry so future runs stay consistent.)
 
    Current migrations: `0000` core tables, `0001` signup tokens,

packages/db/src/migrate.ts

@@ -6,12 +6,26 @@ import { createDb } from "./client.js";
 const HERE = dirname(fileURLToPath(import.meta.url));
 export const MIGRATIONS_DIR = join(HERE, "..", "migrations");
 
-/** Apply all pending Drizzle migrations to the target database. */
+/** App-wide advisory lock id for migrations. Concurrent `db:migrate` runs (e.g.
+ *  api + worker pre-deploy firing off the same push) serialize on it instead of
+ *  racing Drizzle's journal writes. */
+const MIGRATE_LOCK_ID = 0xd1770;
+
+/** Apply all pending Drizzle migrations to the target database. Safe to run
+ *  concurrently: callers serialize on a Postgres advisory lock, and re-running
+ *  applied migrations is a journal no-op. */
 export async function runMigrations(connectionString: string, migrationsFolder = MIGRATIONS_DIR): Promise<void> {
   const { db, pool } = createDb(connectionString);
+  const client = await pool.connect();
+  try {
+    await client.query("SELECT pg_advisory_lock($1)", [MIGRATE_LOCK_ID]);
     try {
       await migrate(db, { migrationsFolder });
     } finally {
+      await client.query("SELECT pg_advisory_unlock($1)", [MIGRATE_LOCK_ID]);
+    }
+  } finally {
+    client.release();
     await pool.end();
   }
 }

railway.json

@@ -0,0 +1,6 @@
+{
+  "$schema": "https://railway.com/railway.schema.json",
+  "deploy": {
+    "preDeployCommand": "npm run db:migrate"
+  }
+}