Brandon Wilson

@njlie Want to open a pull request in https://github.com/interledger/open-payments removing/fixing whatever `access` restriction are causing problems (maybe just use the `uniqueItems`/`maxItems` for now), and update this to use the specs...

An alternative would be to not require auth at all to read Open Payments quotes

:heavy_check_mark: correct

Completing an incoming payment is currently possible via Open Payments: https://docs.openpayments.guide/reference/complete-incoming-payment Rejecting an incoming payment is indirectly possible when incoming payment grant authorization requires RO authorization: - https://github.com/interledger/rafiki/pull/588

Using `apiextensions.k8s.io/v1` crds instead of `v1beta1` might resolve this. I'd noticed invalid extra fields weren't being rejected in k3s for my `v1beta1` crd. I started seeing expected rejections after switching...

> @wilsonianb is the concern here just speed? yeah

I think @DonChangfoot inspired this issue from his initial Rafiki integration attempt. I'm not sure if it'd be preferable for an ASE to poll specific events vs a specific outgoing...

Things are better, but we're not all the way there yet. I've updated the issue description/comment. I believe this is the main problem: https://github.com/interledger/rafiki/issues/443#issuecomment-1220445410

https://github.com/interledger/rafiki/blob/b70323a6abc8b712487891dd7ee4f21b212d125e/packages/backend/src/shared/baseModel.ts#L48-L54 I'm seeing `pagination`'s `first`/`last` being a string (not coerced to `number`) or `undefined` (not set to default when appropriate). `last` as a string causes use to not error both...

- https://github.com/interledger/open-payments/issues/136 I had the thought :point_up_2: could help, but I imagine it would also use `oneOf`...