Identify the cause of the problem between the proxy and the Socket.IO server

[El-Tommy]

I have a Socket.IO server (4.5.2) running with pm2 on port 80.
For HTTPs connections I use an NGINX server, which redirects HTTPS to HTTP and WSS to WS.

It turns out that a few days ago the NGINX server was unable to connect to the Socket.IO server, always returning the HTTP code 504 (Time Out).

Is it possible to identify the reason for this unavailability on the Socket.IO server?

Looking at the pm2 list there was no restart for the application, which means that there was no error in the application that would cause it to restart and go offline.

After this incident I also added the code below, I don't know if it could interfere in the future in detecting the reason for unavailability.
socket.on("error", () => {});

[darrachequesne]

A HTTP 504 error suggests the server did accept the connection, but took too long to respond, This might happen for example when the server is overloaded. Maybe some sync I/O operations?

[El-Tommy]

Looking at the logs I didn't identify any CPU usage above normal or even high LOAD.

One suspicion is that it may have reached the ulimit open file limit. But I don't know how to confirm whether that could have been the case.

[darrachequesne]

Please check https://socket.io/docs/v4/load-testing/