Troubleshooting Web Messaging Latency with On-Premise Edges

Hello everyone. I am an Edge appliance admin and we are currently testing Web Messaging with our on-premise Edge servers. We are seeing significant latency (over three seconds) when a customer sends their first message. I suspect our corporate firewall is performing ‘Deep Packet Inspection’ on the WebSocket traffic, which is slowing down the initial handshake. Has anyone seen this with on-premise Edges, and are there specific firewall ‘Bypass’ rules we should implement for the messaging ports?

Hey Rin85! I am a voice biometrics enthusiast and I deal with these latency issues for our audio streams. You should definitely disable ‘SIP Inspection’ and ‘SSL Decryption’ for any traffic going to the Genesys Cloud cloud-side services. Edges are very sensitive to any delay in the signaling path. If your firewall is trying to ‘Inspect’ the WebSocket, it will almost certainly cause the latency you are seeing.

Greetings. I am an admin for a German company and we have very strict firewall policies. To follow up on Man97, you should also ensure that ‘TCP Keep-Alives’ are enabled on your firewall for the messaging ports (443 and 80). If the firewall closes the connection too early due to inactivity, the customer will see a ‘Reconnecting’ message, which adds several seconds of delay to their next message. We had to increase our timeout to thirty minutes to keep the sessions stable.

Hello! I am a digital channels specialist and I want to add that you should also check the ‘Edge Health’ metrics in the Admin menu. If your on-premise Edges are hitting their memory or CPU limits, they might be slow to process the messaging events. Web Messaging is less resource-intensive than voice, but it still requires some overhead for the session management. If you have many active chats, your Edges might need a hardware upgrade!