WebSocket Ping-Pong Timeout and Connection Closure Failure

[kartikgupta2607]

Is there an existing issue for this?

• [X] I've searched for any related issues and avoided creating a duplicate issue.

Description

Our custom wrapper module built on top of ws handles connection events (ping, pong, errors, message, close) on both server and client sides.

Normal workflow involves:

• Client connects with the server and sends a request.
• Server processes the request and sends a response.
• For long-running requests or multiple requests on the same WebSocket connection, ping-pong calls are sent from the client at regular intervals to ensure the connection remains alive.

Observed Issue:

• Occasionally, a ping call sent from the client is not received at the server end (or at least not logged until this point where we have a ping handler which logs the ping request).
• Consequently, a pong timeout issue occurs on the client end, leading to custom retry logic and eventual closure of the connection if retry also fails.

Failure in Connection Closure:

• After a ping failure, we retry it once, but if it fails again, we attempt to close the connection using our custom close code 3001 (ws.close(3001)).
• However, even the closing handshake fails due to a timeout issue here.

Additional Request

• Is there a way to identify if a connection is stuck on either the server or client side?

Please let me know if you require additional details.

ws version

8.12.1

Node.js Version

Server (18.12.1), Client (14.18.1)

System

No response

Expected result

No response

Actual result

No response

Attachments

No response

[netizen-ais]

The failed pings should indicate network problems in either side, so the "closing handshake" can't complete (as it implies packets between both endpoints)

[lpinca]

Is there a way to identify if a connection is stuck on either the server or client side?

You can capture the traffic with something like tcpdump or Wireshark.

[lpinca]

I'm closing this as answered.