Skip to Content
FastracDocs

Go-live checklist

Everything to check before you switch your integration from the sandbox to production.

Product
APIShipping Engine API
Section
Get started
Reading time
4 min read

Work through this list before you send your first real shipment. Each item links to the page that explains it.

Credentials and hosts

  • Your production secret is in your secret manager, not in code or logs. See API credentials.
  • Production configuration uses https://api.fastrac.example and the production token URL. See Sandbox and production.
  • Sandbox and production settings can't mix: a sandbox token is rejected by production.
  • Your code doesn't call simulate-status in production. It returns 403 there.
  • All calls go from your server. No secret or access token is ever sent to a browser or mobile app.

Access tokens

  • You cache the access token and reuse it until shortly before expires_in runs out.
  • On 401, you get a new token once and retry the call once. If it fails again, you stop and alert.
  • You know who can rotate the key, and how to deploy new secrets quickly. Rotation stops the old secret at once.

See Authentication.

Balance

  • Your production balance is topped up in the dashboard. There is no top-up API, and sandbox balance doesn't carry over.
  • You watch the balance with GET /b2b/user-balance and alert before it runs low. When it runs out, orders fail with 400 Payment failed, but order '…' was created successfully. Reason: Insufficient balance..
  • You can reconcile charges, refunds and adjustments from balance logs. See Reconcile your balance.

Webhooks

  • Your production endpoint is registered with POST /b2b/webhook/setWebhook on the production host. Sandbox registrations don't carry over.
  • The registration response shows challenge.success: true, and GET /b2b/webhook/info shows verificationStatus: "verified". An unverified endpoint receives no events.
  • You verify every signature with the production webhook secret. See Verify signatures.
  • Your endpoint answers with a 2xx within the timeout, then does the work. It ignores duplicates by x-webhook-id, and doesn't rely on event order.
  • A daily job refreshes shipments that are not yet final, in case an event was missed. See Delivery and retries.

Creating orders safely

POST /b2b/orders pays for the shipment and books the courier in one call. There is no idempotency key, and the same shippingRateCode can be booked more than once. A blind retry after a timeout can therefore create a second shipment and charge you twice.

  • You never retry POST /b2b/orders automatically after a timeout or a 5xx.
  • Instead, you first check whether the order was created: call GET /b2b/shipments with search set to the receiver's phone number and startDate set to just before your request. The search field also matches shipment codes, order codes, airway bills and receiver names.
  • If you find a match, confirm it with GET /b2b/shipments/{code}/details, whose shippingRateCode should match yours. Only if nothing is found do you request a fresh rate and book again.
  • You save the order code and every shipment code as soon as you receive them.
  • On 400 Payment failed, but order 'FSO-…' was created successfully., you don't retry with the same request. The order is not booked and any charge is refunded. Fix the cause, then book again.

Error handling

  • You show the message of 400 errors in your logs, and don't retry them unchanged.
  • Expired rates (Shipping rate code '…' have expired.) lead to a new rate request and, if the price changed, a new confirmation from the customer.
  • 422 errors (balance, wrong state) are surfaced to an operator, not retried.
  • Only 5xx and network errors on read calls are retried, with exponential backoff.

See Errors and Handle errors and retries.

Monitoring

  • You log the HTTP status, message and your own request ID for every call, without logging secrets or tokens.
  • You alert on a rise in 401, 403, 422 and 5xx responses.
  • You alert when your webhook endpoint starts failing. After many failures in a row, delivery pauses for a while and events sent during the pause are lost.
  • You track shipments stuck in one status for longer than the courier's estimate, and PROBLEM_SHIPMENT.

Getting help

If something isn't working, contact Fastrac support through the support page in the dashboard, or your account manager. Include:

  • the environment (sandbox or production);
  • the order code (FSO-…), shipment code (SHP-…) or airway bill;
  • the endpoint, the time of the request with its time zone, the HTTP status and the full error message.

Never send your API secret or webhook secret to support.

Was this page helpful?

On this page