It’s down… now what?

Step 0 — Don’t panic

Open /api/health first. It tells you whether:

• the app can read config
• the DB is reachable
• migrations are pending
• the worker heartbeat is recent

Step 1 — Is Postgres up?

Symptoms:

• /api/health says db.ok=false

Fix:

• Confirm DATABASE_URL / config credentials
• Confirm Postgres is reachable from the server
• Confirm firewall / networking

Step 2 — Are migrations pending?

Symptoms:

• /api/health shows pending migrations > 0

Fix:

npm run db:migrate

Step 3 — Is the worker running?

Symptoms:

• Crawlers never run
• No recent worker.heartbeat in /api/health

Fix:

• Restart worker process
• Ensure you are running npm run worker as a long-lived process

Step 4 — Are crawls failing (HTTP errors)?

Symptoms:

• Runs show ERROR

Fix:

• Lower RPS / concurrency
• Some portals throttle. Keep it polite.
• Verify the portal does not require login.

Step 5 — Worst case: rebuild the registry (testing)

If this is a test environment:

• /app/admin → Reset registry DB schema
• or npm run db:reset

If this is production, restore from backups.