#Will-webhook
wet sequoia
How often will the reties happen?
E.g. We're planning on putting our API into "maintenance" mode for ~1hour due to a huge DB change. I want to make sure it won't affect Stripe webhooks.
heavy bison
My suggestion is still giving 200 back to Stripe, but record internally the event to resume later
wet sequoia
The docs only says If Stripe doesn’t quickly receive a 2xx response status code for an event, we mark the event as failed and stop trying to send it to your endpoint. Do you know what quickly means in this context?
record internally the event to resume later
I'm not fully sure how that's supposed to work, could you elaborate further please?
heavy bison
quickly just mean we can receive the response without a timeout error
My previous suggestion is, you likely build a balancer standing before your "real" logical server. When you take your logical server down for maintenance, let's say 1 week, you still let the balancer responding 200 to Stripe without actually forwarding to your logical server
You would need to queue up the events in that 1 week somewhere for your logical server to resume consuming it later
wet sequoia
Ok, but if we're going to be down for more or less an hour, I guess it shouldn't be a problem?
Or would you still recommend building that 200 responding balancer, even if we're down only for an hour?
heavy bison
More info above
1 hour should be fine!
multiple days in a row would be problematic
wet sequoia
Got it, thank you! Have a lovely day