In the stripe documentation, it says:
So in this case, the checkout page goes to the success or failed page on my frontend.
I use the backend to track the payment status so that we can monitor the transactions in the admin portal, and the above approach seems dangerous to me.
When checkout is successful, it redirects the window to the success url. This means I have to call the backend API in the success page to update the payment status. However, the stripe is the source of truth about the payment status, and the status update on DB should come from Stripe, not come from a frontend page. At the very minimum, if a user refreshes the success page, it would have called the API again and again which is bad. Also, it is about "a user says I paid successfully" v.s. "Stripe says they paid successfully".
I tried the Stripe webhooks, but in the webhook data object, there is no information that I can use to link it to the sessionId that is generated from creating the checkout session, but the session id is the only tracking id I can get from Stripe about a payment.
What's the best practice, if Checkout is the only solution, to securely update my database?
You have 2 options:
Rely on webhooks. The checkout.session.completed
event will describe a Checkout Session which contains its ID, which you hopefully saved when you created the Session earlier so you can link the two together.
Retrieve the session ID from the success URL once the payment is complete and retrieve the Session on your server, then check the Session's payment_status
. This way your server can verify if the payment was actually completed or if someone just managed to guess the URL of your success page.
Stripe doesn't recommend only doing option 2, as it's very possible that users close the browser tab or window before the redirect to your success page can happen, resulting in a possible loss of payment confirmation. You should always use webhooks instead to guarantee your purchase fulfillment logic correctly fires.