Experiencing intermittent SIP 408 Request Timeout errors specifically within our Genesys Cloud Architect flows on the AP-1 region BYOC trunk. The issue manifests when callers navigate through a multi-level IVR menu structure, particularly after selecting an option that triggers a downstream API interaction via the “Invoke REST” block.
The trunk configuration remains stable with valid SIP registration, and latency metrics to the carrier gateway in Singapore show nominal values (<15ms). However, once the flow executes the API call and attempts to return control to the “Play Prompt” or “Collect Input” block, the session drops with a 408 error from the carrier side. This behavior is not observed on voice-only trunks or within the US2 region, suggesting a specific quirk with how the AP-1 gateway handles prolonged INVITE states during asynchronous processing.
Current environment details:
- Genesys Cloud Platform Version: 2024-08
- BYOC Trunk: AP-1 Region, 15 concurrent sessions active
- Architect Flow: Standard IVR with 3-level menu, API timeout set to 5000ms
- Error: SIP 408 Request Timeout logged in the Conversation History
Has anyone encountered similar timeout behaviors when integrating REST API calls within IVR flows on AP-1 BYOC trunks? The carrier has confirmed no firewall drops on their end, and the SIP logs from Genesys indicate the timeout originates after the API response is received but before the next media leg is established. Investigating if there is a specific timeout configuration in the Architect flow or trunk settings that needs adjustment for this region. Any insights on carrier-specific quirks or failover logic adjustments that might mitigate this would be greatly appreciated.