https://railway.com/project/6d0f799e-be59-4388-899d-f00456f30667/service/68f526be-f01d-4a18-9978-cb549b4e5d7a?environmentId=5ef05232-a59c-4267-a86d-a657f4d3fab4

• nest js

• Prisma

• Connects to railway postgres

Hello,

I would love to help if I could, but we don't have any visibility into what your code is or is not doing, so there's nothing we could help you with on our side.

hey, i started having this same issue yesterday, however with Bun and not node. i use Prisma and Postgres on Railway. my logs seem to indicate that there is a connection failure to the Postgres db and then right after that i get segfaults / double frees and crashes. there is really nothing that changed in my app that has been completely stable for months that i think would have caused this.

i seem to get these kind of logs on my postgres container seconds before the crash
2025-08-01 16:08:21.379 UTC [23326] LOG: could not receive data from client: Connection reset by peer
2025-08-01 16:08:21.379 UTC [23323] LOG: could not receive data from client: Connection reset by peer
2025-08-01 16:08:21.379 UTC [23325] LOG: could not accept SSL connection: EOF detected
2025-08-01 16:08:21.379 UTC [23327] LOG: could not accept SSL connection: EOF detected
2025-08-01 16:08:21.379 UTC [23322] LOG: could not receive data from client: Connection reset by peer
2025-08-01 16:08:21.379 UTC [23324] LOG: could not accept SSL connection: EOF detected
2025-08-01 16:08:21.380 UTC [23314] LOG: could not receive data from client: Connection reset by peer

We kinda identified our issue to be related to prisma. Not fully sure we have solved it but related to: https://github.com/prisma/prisma/issues/27785

Thanks @Brody I know this one might not be on railway. but I couldn't find anything that changed on our end. It seems to be related to an libopenssl update

sounds a lot like this is the issue, thanks. railway was definitely not my first suspicion for the cause but i started to wonder when i noticed connection issues preceding segfaults

I had zero connection issues preceding segfault. It took all day and everybody to identify this

I can't blame y'all for looking at Railway for this; it's a fair assumption when nothing changes in your codebase, but I'm glad you found the issue!

!s