WebRTC causes FreeSWITCH to not provide service

Question

I am using two FreeSWITCH clusters, where the difference between them is that Cluster A uses TCP protocol for SIP and Cluster B uses webRTC protocol for SIP. Recently, there have been several instances where one of the FreeSWITCH servers in Cluster B suddenly stops providing services, but has never occurred in Cluster A. Therefore, I suspect that webRTC may be causing FreeSWITCH to stop providing services. Here are more details about the situation:

agent & call: Each FreeSWITCH instance has approximately 100 agents registered, and there are around 20,000 outgoing calls made per day. Frequency of issue: Occurs approximately once a week Symptoms of issue:

1. Ongoing calls are not affected;
2. New outgoing calls cannot be originate;
3. Users are unable to register with the affected FreeSWITCH server;
4. The CPU and memory usage of the 2C4G configuration are both very low;
5. There are many network connections in CLOSE_WAIT state;
6. The version of FreeSWITCH being used is 1.10.9, and the jsSIP version is 3.9.4.

Please advise on how I can troubleshoot this issue;

Restarting FreeSWITCH restores normal operation.

Answer 1

Sounds like you may be experiencing this Freeswitch bug. You might consider not using Freeswitch as the wss server. Freeswitch seems to have some pretty serious issues with websockets, especially secure websockets, and they don't seem interested in fixing it. although that thread is only a few months old, projects that use Freeswitch like BigBlueButton have been working around this for a long time.

This issue from BigBlueButton talks about how to use nginx for the TLS termination. This requires some client modification to work, as Freeswitch expects a