We create PaymentIntents (with capture_method=manual in case that matters) in our iOS/Android apps when the user places an order.
We send the order to the connected venue once the charge.succeeded webhook fires. If this doesn't happen within a couple of minutes, we expired the placed order on our side.
So interestingly 2 out of 10 times we don't get this webhook to fire.
Im wondering if it's actually wise to listen to this webhook in order to decide if we send the order to the connected account's venue or not or if there is a better way to determine that the payment will actually work once we try to capture it.
Webhooks are the recommended way for getting a payment intent's status, but you can also use the API to get its status.
A quote from the Stripe docs:
It is technically possible to use polling instead of webhooks to
monitor for changes caused by asynchronous operations—repeatedly
retrieving a PaymentIntent so that you can check its status—but this
is markedly less reliable and may pose challenges if used at scale.
Stripe enforces rate limiting on API requests, so exercise caution
should you decide to use polling.
In your case, I'd recommend waiting for the webhook and then after a few minutes, call either the PaymentIntents API or the Charges API if you haven't received the webhook yet.
There's likely something else going on here, so I'd suggest you reach out to Stripe - webhooks should fire all the time, and it's a really really rare occurrence that they wouldn't.
Related
I am building an app to sell single item product (i.e, each kind of products listed on my platform only has a single item).
(this part has been done, and won't change) I built an in-house backend having the Rest API POST -D {"buyer_email": "abc#example.com"} url/items/{itemID}, let's call it transaction_call, which will make sure once a customer succeed the POST operation, his contact info is recorded into my backend as the successful buyer; and all other customers will fail to buy that item (at API level, transaction_call return 4xx error) because my platform can only sell one item for that product;
(this is the step that my current question is about) I am trying to use Stripe as my payment system on this platform.
I really want to integrate with Stripe as simple as possible (as I understand Stripe Checkout is the most simple / out-of-box way to implement payment). However, I am not sure if Stripe Checkout can achieve the above functionality correctly. Since the problem is a two-step problem, here is the potential issue I may run into:
Let's say, two customers A, B, get to my platform at 10:00am, both of them start purchasing process for a product, Item_a
If my system interact / call the Stripe Checkout first as the first step then call the transaction_call, here could be the problem:
A's Stripe call hits the Stripe server at 10:00:01am, and A's buying call hits my backend at 10:00:02am;
B's Stripe call hits the Stripe server at 10:00:01am, and A's buying call hits my backend at 10:00:03am;
in this way, we have already charged B but he really did not get the item
If my system calls the transaction_call first, and only if transaction call succeeds then it interacts / calls the Stripe Checkout, then
A's transaction_call succeed at 10:00:01am, but he for some reason decided not to pay (not click confirm button on the Stripe Checkout UI)
In this way, my system fails to sell the item to other buyers.
My question is whether the above reasoning process is correct, and whether I could somehow use Stripe Checkout to achieve what I am doing.
Maybe I have implement the payment functionality using Stripe Intent API to build a workflow-based payment, which I assume will be much more complex, if the Stripe Checkout way (simple wayO is really not possible.
From what I understand you have a potential race problem, where the item you're selling is very limited in quantity and you want to make sure that you can correctly notify users if it's out of stock or already spoken for.
For your first scenario, the simple solution is only invoke Stripe's API on your backend when you've received the transaction_call. For instance, you'd only create the Checkout Session once your system has identified that the item is still available. You'd then "lock" the item so that when B attempts to purchase you can immediately return an error instead of creating a payment via Stripe's API. The logic on who to charge (basically who initiated the checkout process first) in the case of a tie would then be for you to implement in your transaction_call rather than on Stripe's side.
The second scenario is a little tricker, as Checkout Sessions can't be cancelled once you create them. They automatically cancel themselves after 24 hours if no payment is made, but I doubt that you'd want B to have to wait that long if A abandons the payment flow.
Instead I think you should look at implementing a PaymentIntents integration, where you can more finely control the flow.
Your flow for scenario 2 could be:
A begins the checkout process, create a PaymentIntent on the backend, "lock" the item and start a timer
The timer (which you'd ideally show to your user) times out after N minutes if A doesn't pay
Cancel the PaymentIntent on your backend and remove the lock
B can now attempt to pay for the item, upon where you restart the process
I'm updating a webapp to Stripes SCA ready flow with PaymentIntent.
So far I have working (on my local test server):
Generate Intent on frontend and pass secret to form
Use Elements to collect card into
Use handleCardPayment to create the charge
Now here's the part where I am unsure. The handleCardPayment responses all seem to indicate a succeeded event, but the documentation warns to not use this repsonse, but instead wait for the Webhook response and only then fulfill customer orders.
Step 5: Asynchronously fulfill the customer’s order
You can use the PaymentIntent returned by Stripe.js to provide
immediate feedback to your customers when the payment completes on the
client. However, your integration should not attempt to handle order
fulfillment on the client side because it is possible for customers to
leave the page after payment is complete but before the fulfillment
process initiates. Instead, you will need to handle asynchronous
events in order to be notified and drive fulfillment when the payment
succeeds. Documentation
So far so good, I've set up test webhooks and tunneling through ngrok I can actually receive the paymentIntent from the Stripe webhook.
Now, my question comes at this point, where the Stripe documentation ends. How should I deal with the UI from the point of the "Pay" button being pressed, and how do I in the frontend detect that the webhook has been triggered?
I am wondering if I should poll my own server, which in turn retrieves a database result that indicated if the webhook for this order has been received? Or what is a reasonable way to deal with this, technically and from an UX perspective?
Any pointers?
I just implemented this, and I decided to poll my own server for the update and ask the user to wait. The webhook marks our internal representation as "paid", so we don't need to poll Stripe. If the webhook doesn't come within 30 seconds, we tell the user that it's ok to leave the page, and we'll email them the result.
Technically, these webhooks can take up to 7 days to come in, so that's why Stripe doesn't want you to have the user wait. In reality, it almost always comes in within 5 seconds, and I would rather just have the user wait and see a final confirmation in the same session.
The Square documentation for updating webhook events shows this URL format: PUT /v1/{location_id}/webhooks. However, creating a webhook event listener for every merchant location could be a lot of separate API requests, and it would be far easier to use the merchant_id instead of the location_id (even though this is not documented) and make one request for each merchant.
Attempting to do this actually works - when I PUT /v1/{merchant_id}/webhooks the webhook is saved in Square and transactions for any of that merchant's locations successfully send the webhook.
My question is, since this is undocumented (although it works) is it safe to rely on this approach?
While it may work currently, since it's undocumented, the behavior may change in the future and cause unintended side-effects. I strongly encourage you to follow the current documentation for subscribing to webhooks.
I would like to integrate my web application with the Square POS.
The goal would be to be notified each time a transaction (sale/refund/etc) is processed by Square for an account so that I can update inventory levels etc, ultimately so I can update inventory levels as transactions occur.
From what I can tell, it seems that the Square API's seem to be designed around my application initiating the transaction, then handing off to Square to process the payment. I simply want to be notified that a transaction has happened so that I can update inventory.
Is it possible to do this? Or is the Square API just for processing payments?
edit: After some more reading, I still haven't found a webhook to be notified, but it looks like I can ListTransactions, and RetrieveTransaction, so if I poll I should be ok.
You’re correct. Square’s API Webhooks will be what you’ll use to be notified each time a transaction is created or updated. We have a quick setup guide available in Square’s Developer Doc (https://docs.connect.squareup.com/api/connect/v1/?q=webhooks#setupwebhooks).
The PAYMENT_UPDATED webhook will alert you every time a payment is made, so that you can update your inventory.
My fulfillment needs to do a lot of processing after receiving a certain request from Google Action/API.AI and the default response timeout is 5s.
https://developers.google.com/actions/components/fulfillment#nodejs
Is there any way I could send a delayed response or send a POST request after the results are ready?
The short answer is no - you must respond within 5 seconds, and there is no way to send a notification back through the Assistant at this time.
The slightly longer answer is that we know notifications are coming - but we don't know if there will be an API for them. There have been rumors about other ways that may be coming that allow us to work around the 5 second limit.
The even longer answer is that, if you are using Action Transactions (ie - allowing the user to purchase or reserve something) you can send updates after the fact. However, Transactions are still in developer preview and don't work on all surfaces (they don't work on Google Home at all, for example).