#tinyzimmer-webhook-http2
1 messages ยท Page 1 of 1 (latest)
A little more context - the backend on the load balancer is using h2c (to support gRPC connections for other stuff). The fix on my end was just a separate load balancer talking HTTP/1.1.
๐ Quick question for you - was the 464 an error coming from Stripe? Do you have an example event ID I can take a look at?
Nah the 464 came from my AWS ALB
Yea I'll grab some IDs for you
pm_1JXwN2IyDFbcdgGyrOIHKS04 - is a good one - you can see the retry worked against the new LB
let me know if you need more - it was broken for a while ๐
That's the PM ID - do you ahve the event ID? (evt_xxx)?
oops
evt_1KPvoUIyDFbcdgGye7brNLkV
think this is the one i resent evt_1KPvoUIyDFbcdgGye7brNLkV
๐
looking now!
I'm not finding any specific mention of why we don't support HTTP2, but I can flag this as feedback to add support for it in the future
Great thanks! To be fair it was an assumption. I ran into similar issues with another webhook service that was solved the same way. And before I switched to AWS WAFs I was using cloudflare and it couldn't handle the HTTP2 backend either.
The issue CF side was better documented, but they all had the same 464 symptom
I also looked through your docs for specific mention, couldn't find any so I came here.
Yeah, this just isn't very well documented on our end right now - but hopefully I can get the fixed ๐
Grazi ๐
If you care for more reading - a thread of angry AWS customers experiencing the same issue I did with other webhook services (e.g. twillio) https://forums.aws.amazon.com/message.jspa?messageID=985494
It's more nuanced than you just "not supporting HTTP2" probably. I mean if possible the requester should catch the offer to upgrade and take it. But really it's AWS dropping the ball here imo.