3 months ago
https://station.railway.com/questions/edge-network-broke-stripe-webhooks-750dd85b
Continuing from this solved thread, I'm also experiencing Stripe timeouts but my port is set correctly. The services were deployed today so I've never had a previously working version of it.
Stripe CLI returns 405 errors
uptrends.com/tools/uptime shows 405 errors
SSL Report gives my .up.railway.app URL a perfect report
Stripe dashboard (where the real events will be sent from) has HTTP Status code "Timed out"
1) I confirmed the webhook secret is correct
2) I confirmed the URL is correct
3) I have Public Networking set to Port 8000 (the one Railway detected)
4) Before deploying on Railway, my Stripe webhook code worked perfectly
The "Metal Edge Broke Stripe Webhooks" help thread is the only thing I can find about this issue and it was closed 6 hours ago. Can we revisit and try to find a different solution?
Project ID 23102505-86b3-4c1e-9549-4642b8e6ab93
3 Replies
I don't think this is solved, but downgrading my .up.railway.app public networking domain from metal edge fixed the issue.
I want to use metal edge so I'm hoping there's a better solution
Can confirm that metal edge networks seem to be unstable, I had similar issues with MinIO https://discord.com/channels/713503345364697088/1348113300981813320
Hoping this gets fixed - I spent a while on this because I thought some configuration was wrong in the service itself
Hey, I was the OP from the original thread.
I closed it because it was working at that time, but only briefly like only for 20 minutes 😦
I should have reopened the thread, sorry.
But I can confirm that it's still broken, I also ended up downgrading Metal Edge
In addition to the timing out issue, my backend has 2 public networking URLs:
.up.railway.app (set to old servers)
api..com (set to new servers)
When my frontend's API_URL was set to api..com the requests were taking up to 10x longer than when it's set to .up.railway.app. Whether this is because of the Metal servers or the DNS forwarding api..com to my backend I'm not sure
But that could contribute to Stripe timing out. I read somewhere Stripe expects a response in 10 seconds so if my code's not sending a response in <1 second then the 10x slowdown may be causing it to time out
Confirming that downgrading to GCP has fixed the stripe webhook timeout issue
3 months ago
hey @here - with the help of one of our users we belive we have found and fixed an issue with the metal edge proxy, would any of you be so kind to test it again?
Hi @Brody, it still doesn't work.
Enabling the Metal Edge immediately caused Stripe Webhook to fail.
3 months ago
thank you for testing, we will be going back to the drawing board
3 months ago
leaving a message so I can get ping if this gets fixed
3 months ago
Same problem for me
3 months ago
Same issue for me. Had to downgrade to allow it to work again.
3 months ago
Same issue here. Downgrading fixed it.
3 months ago
Hello @here - We have made an additional change, and in our testing stripe can now reach Railway hosted services with the Metal Edge.
We would appreciate it if you could verify if it is working now?
I just tried testing the webhook again, seem to work fine now!
I'll monitor it until the next few days.
3 months ago
awesome!!
3 months ago
just curiosity: what was happening to cause this?
3 months ago
we updated the advertised MSS to 1440
3 months ago
<:lfg:786867302888505364>
3 months ago
!s
Status changed to Solved brody • 3 months ago