What is the standard approach to handle SIP 408 Request Timeout errors when our primary carrier fails over to the secondary leg on our AP-SE-2 BYOC trunks? We manage 15 trunks across the region, and recently the failover logic has become erratic. When the primary SIP registration drops, the outbound call attempts to bridge to the secondary carrier, but Genesys Edge consistently returns a 408 before the secondary leg fully establishes. This results in dropped calls rather than successful handoffs.
The issue appears specific to the Singapore edge nodes. We have verified that the secondary carrier accepts the INVITE successfully if called directly, but the edge timeout seems too aggressive during the transition window. Our current SIP configuration uses the default keep-alive intervals, and we are running SDK v2.14.0. The analytics report shows a spike in call_failed events with reason_code set to network_error precisely when these timeouts occur.
- Increased the
invite_timeoutandresponse_timeoutvalues in the trunk configuration to 15 seconds, but the 408 persists at the 12-second mark. - Verified that the secondary carrier’s SIP proxy is reachable via ping and telnet from the edge IP range, confirming no network-layer blockage exists.