Hey everyone, I’ve run into a really strange issue with our predictive routing weights during peak hours in Singapore. The system fails to adjust the weight decay for calls routed through our secondary BYOC trunks when the primary carrier hits its concurrency limit. This results in a 408 Request Timeout on the outbound leg, even though the SIP INVITE was accepted by the carrier gateway.
We are using Genesys Cloud API v3.1.4 for trunk configuration. The failover logic triggers correctly at the trunk level, but the predictive routing engine seems to cache the initial weight calculation. Is there a known issue with weight recalculation latency during carrier failover events in the APAC region?