Could someone clarify why the SIP trunk registration keeps failing with a 408 Request Timeout? We migrated from Zendesk Talk, where static IP whitelisting was sufficient, but Genesys Cloud seems to require strict NAT traversal settings that are breaking our inbound flow. The logs show the INVITE packet is dropped before reaching the media server. Is this a known issue with the current EU-west-1 edge deployment?
Yep, this is a known issue… The 408 usually indicates the SIP proxy is dropping the packet due to high concurrent registration attempts, not just NAT settings.
Check your JMeter script if you are load testing. Ensure you aren’t exceeding the WebSocket connection limits for the platform API before the SIP handshake completes.
Yep, this is a known issue. while the 408 timeout often points to proxy congestion, the root cause here is likely the sip trunk registration timeout settings conflicting with your edge latency. the default registration interval is too aggressive for the eu-west-1 edge during peak loads.
try adjusting the sip trunk configuration in the admin console. increase the registration timeout to 30 seconds and set the expiration to 3600 seconds. this reduces the frequency of re-registration packets hitting the proxy.
{
"registrationTimeout": 30,
"registrationExpiration": 3600
}
also, ensure your firewall allows outbound traffic on ports 5060-5091 for udp/tcp. if the nat traversal settings are strict, the keep-alive packets might be dropped, triggering the timeout. check the sip trace logs for any 407 proxy auth failures too. this usually stabilizes the connection without needing to tweak the edge deployment settings directly.