15 days ago
We have been facing intermittent 502 connectivity issues, where the backend pod would stop processing requests. This is causing constant intermittent issues on production and it doesnt relate to any increase in load.
3 of the most recent instances
Jun 20, 2026, 10:29:49 PM GMT+8, deployment hash: 0bb948af
(these 2 are happening live and back to back)
Jun 22, 2026, 2:41:03 PM GMT+8, deployment hash: 538a63cc
hash: ccc24231 (like 12 mins later, died again)
Attachments
1 Replies
Has been resolved by angelo the goat (https://station.railway.com/questions/networking-outbound-connection-problem-b3be7558)
tldr; move service to use ha static ip address, issue seems to be caused by dynamic ip address getting evicted/intermittent internal nat routing being unavailable.
Status changed to Solved brody • 15 days ago