Hello!

We're acknowledging your issue and attaching a ticket to this thread.

We don't have an ETA for it, but, our engineering team will take a look and you will be updated as we update the ticket.

Please reply to this thread if you have any questions!

Hello,

That would be around the time we cycle the proxies. We are aware of this issue and will be looking into mitigations.

Best,
Brody

Appreciate the quick response - anything we can do in the short term? This is impacting our ability to serve customers.

Even just shifting the proxy cycling outside of core business hours, while maybe not a perfect solution, would really help reduce business disruption.

Shifting the cycle time of the proxies wouldn't be a solution, unfortunately. It may mitigate the issue for you, but it would just move it to someone else's business hours, so we will opt to keep the time it has always been set to.

A solution could be to deploy the Cloudflare tunnel, thus bypassing our proxy.

https://railway.com/deploy/cf-tunnel

Again, I want to acknowledge that this situation in general is not ideal, and I apologize for that.

Hey, small update here. As mentioned, we would restart the proxy every day. We have changed that to be a little more intelligent with those restarts. Now, restarts are performed only when needed. In practice, that means fewer restarts.

🛠️ The ticket Websocket traffic disruptions has been marked as triage.

🛠️ The ticket Websocket traffic disruptions has been marked as backlog.

❌ The ticket Websocket traffic disruptions has been marked as canceled.

Hey! Still seeing a massive amount of 1006 disconnects and other networking-related issues. It's been especially rough the past ~two days. Any updates on this? Has something changed recently?

Anything that can be done to mitigate this would be extremely helpful - this is materially impacting our business.

We've been pushing updates to our TCP proxy for some upcoming features, so the increased restarts you're seeing recently are expected as part of that work. For resilience, we'd recommend having your clients hold multiple WebSocket connections open and multiplexing data over them. That way, each connection will be cycled at a different time, and the redundancy ensures your clients don't drop an entire data transmission when one connection closes.