Openfoodnetwork: Multiple successful stripe payments captured for the same order

[luisramos0]

Description

Most of us are not convinced that #5663 fixed #5662 so I think we need this issue to track the problem in Katuma.

I believe #5703 is a new attempt at fixing this problem. Right @Matt-Yorkley

Expected Behavior

Actual Behaviour

Steps to Reproduce

1.
2.
3.
4.

Animated Gif/Screenshot

Workaround

Severity

Your Environment

• Version used:
• Browser name and version:
• Operating System and version (desktop or mobile):

Possible Fix

[Matt-Yorkley]

Yes :+1:

If you look very carefully through the erorrs here: https://app.bugsnag.com/katuma/katuma/errors/5ef235fc2dfe230018a3c3b8?filters[event.since][0][type]=eq&filters[event.since][0][value]=7d&filters[error.status][0][type]=eq&filters[error.status][0][value]=open it's obvious there are some very serious problems, and multiple checkout submissions _per request_ seems to be one of them. I haven't quite figured out why that is happening, but I think the messy transaction rollback is the crux of it.

[sauloperez]

For the record, Eugeni successfully placed an order with the affected hub but all the timestamps are there. You migh want to check it out @Matt-Yorkley . It's the number R443016306.

See here the options he selected

[Matt-Yorkley]

With the "save default shipping address" option, did they have a default address saved on that account before placing this order?

Maybe removing the timestamp constraints was alleviating the worse bits of the problem?

[sauloperez]

With the "save default shipping address" option, did they have a default address saved on that account before placing this order?

We'll need to ask Eugeni

Maybe removing the timestamp constraints was alleviating the worse bits of the problem?

I assume so

[cillian]

Could the Stripe/ActiveMerchant idempotency_key option help here too when capturing payments?

[Matt-Yorkley]

Looks like this is fixed by #5703 but we can keep an eye on it.

[luisramos0]

Thanks @cillian for that. I think that's a great idea. We have a strategy in place already in #5703 and also #5678 but that is definetely something we should consider if we bump into this issue again.
https://stripe.com/docs/api/idempotent_requests
ping @sauloperez fyi

[sauloperez]

Yep, thanks a lot @cillian . Sorry, I said nothing but I kept that comment on my things to review. The name alone sounds pretty good. Do you have any experience with it?

[Matt-Yorkley]

Maybe we can make a new issue to look at idempotent requests with Stripe, and close this one? It seems like this issue is resolved...

[luisramos0]

I think we can close this one yes.
Re the new issue for idempotent requests, yes we can create a tech debt issue.

[cillian]

@sauloperez no problem. A bit, so I could try do a spike on it at some point maybe.