#tobias-subscription-usagerecord
1 messages · Page 1 of 1 (latest)
Hi 👋
You can disable automatic advancement of the Invoice on a per Invoice basis but you cannot prevent the subscription from creating a new invoice for each billing period.
What are you actually trying to do here?
So, I'm trying to add 'complicated' usages to a subscription, where we measure a timespan something ran that we bill for.
Either we have to add the usages(as I understand it) while the subscription is still going, or we have to keep the invoices in draft mode, while we add a invoice draft item, to the subscription for the usages that happened during the period.
And as I read the docs, https://stripe.com/docs/invoicing/integration/automatic-advancement-collection
It is not 100% clear to me, if we can disable the auto_advance 'ment completely, or if we have to do it on a per invoice basis like you mentioned.
I would recommend using usage records to account for usage during a billing period and then invoice for that usage. That is what our usage-based recurring pricing scheme is designed for. https://stripe.com/docs/products-prices/pricing-models#usage-based-pricing
Lets say I calculate a usage, and add it to the ongoing subscription, wouldn't I risk, e.g. on a five minute interval that I loose the last 5 minutes of actions?
I know there is a clock drift period, but it assumes the calculation can be done in five minutes.
Sorry for the delay, things got really busy there for a sec
I'm not clear on exactly what the concern is here. Are you worried that your integration would not be able to perform updates to usage records fast enough to capture the usage of your customers?
That is exactly right.
If you imagine a tier where the price is 4 usd for 20 users, but 4000 usd for 21 users, the difference between losing a single recorded user in the period could hurt.
This is exagerated ofc, but im one of those people, but I kinda see that this isn't going to be apples to apples, and I am okay with that.
Do we have any guidence on how we are allowed to hit the api, no more than 100 times per second I think I read.
For live mode requests the limit is 100 requests per second. In test mode it's 25 rps
But if you are updating usage every 5 minutes that should be fine. What kind of usage cadence would you be tracking internally?