Hi Railway team,
I’m encountering a persistent issue with my n8n environment deployed on Railway.

Since today, the n8n editor instantly shows “Connection lost” and never reconnects.
Here’s what I’ve already tried:

• Fully redeployed the service

• Created a fresh n8n environment from scratch

• Restarted all linked services (Postgres, Redis if applicable)

• Verified environment variables

• Tried different browsers and devices

The problem remains exactly the same: the n8n UI loads, but the editor loses connection immediately and all workflows appear unavailable.

Given what the n8n community reports, this issue could be caused by one of the following:

• WebSocket/SSE not working correctly behind the proxy

• The Origin header being modified by the hosting layer

• A reverse proxy not forwarding the upgrade request

• A version-related compatibility issue specific to Railway’s environment

• A networking or resource isolation problem on the container

Could you please check:

1. Whether WebSockets are fully supported and correctly forwarded for my deployment

2. If anything in the proxy/CDN layer could be interrupting SSE/WebSocket connections

3. Whether my service logs show restarts, resource limits, or blocked connections

4. If you recommend a specific configuration (env var like N8N_PUSH_BACKEND=sse or others) for stable n8n deployments on Railway

Here is my service URL and ID (if needed I can provide logs):

Service: n8n
Issue: Persistent “Connection lost” on every deployment

Thanks a lot for your help this issue makes the editor completely inaccessible on my side.