Could someone explain why our predictive outbound campaigns are consistently stalling with 503 Service Unavailable errors when the system attempts to route calls through our secondary BYOC trunks in the Asia Pacific region? We have configured a failover logic that shifts traffic from our primary trunk to the secondary one when the primary hits its concurrent session limit, but the transition is causing significant latency and eventual timeouts. The campaign logs show that the predictive engine successfully identifies the need to switch trunks, but the SIP INVITE requests to the secondary carrier are being rejected with a 503 error shortly after the failover trigger is activated. This issue is particularly problematic during peak hours in the Asia/Singapore timezone, where call volumes are highest. We have verified that the SIP credentials for the secondary trunk are correct and that the trunk is registered properly in the Genesys Cloud portal. The outbound routing rules are set to prioritize the primary trunk and fall back to the secondary trunk only when the primary is at capacity. However, the failover mechanism seems to be introducing a delay that causes the predictive dialer to abandon the call attempt before the secondary trunk can process the request. We are using the Genesys Cloud SDK v1.0.20 to monitor the campaign performance, and the metrics indicate a sharp increase in abandoned calls during the failover period. The error messages in the logs do not provide much detail beyond the 503 status code, which makes it difficult to pinpoint the exact cause of the issue. We have also checked the carrier-specific quirks and ensured that the secondary carrier supports the required SIP headers and codecs. Despite these efforts, the problem persists, and we are unable to maintain the desired service levels for our outbound campaigns. Any insights into how to optimize the failover logic or troubleshoot the 503 errors would be greatly appreciated. We are considering adjusting the failover thresholds or implementing a different routing strategy to mitigate the impact on call delivery.
I’d recommend looking at at the trunk failover configuration settings in the Genesys Cloud admin portal, specifically the health check intervals and the retry logic for SIP trunks. When the primary trunk hits its concurrent session limit, the system needs to detect this state change quickly and switch to the secondary trunk without significant delay. Ensure that the health checks are configured to detect trunk failures or capacity limits promptly, and that the retry logic allows for a seamless transition to the secondary trunk. Additionally, check the SIP trunk settings in the Asia Pacific region to ensure they are correctly configured for failover and that there are no regional restrictions or latency issues affecting the transition. Review the campaign logs for specific error messages related to the 503 Service Unavailable errors and adjust the predictive dialer settings to handle trunk failovers more gracefully, such as increasing the retry interval or adjusting the predictive model to account for potential trunk delays.