Trying to understand the exact interaction between Architect flow timeouts and SIP registration states on our 15 BYOC trunks. We are seeing a specific pattern in the Singapore region where outbound calls fail with a 408 Request Timeout from the carrier, but the Architect flow continues to wait for 30 seconds before failing over.
The SIP trace shows the 408 arriving within 5 seconds, yet the flow logic does not trigger the error handling path immediately. Is there a specific configuration in the BYOC trunk settings that delays the signal propagation to the Architect engine? We have verified that the timeout settings in the outbound routing policy are set to 10 seconds, which should be sufficient.
The environment is running the latest patch for the Singapore edge nodes. We are not seeing this behavior on the Tokyo trunks, which suggests a regional routing quirk or a carrier-specific SIP stack issue. Any insights on how to force an immediate failover upon receiving a 408 without waiting for the full Architect timeout cycle? We need to reduce the perceived latency for our agents.
Thank you for your help.