What is the reason the trunk utilization percentages returned by the /api/v2/analytics/conversations/summary endpoint do not align with the raw SIP registration logs for our AP-SE-2 BYOC trunks? We are running SDK v2.14.0 and have fifteen trunks configured with specific carrier failover logic. When querying for outboundCarrierFailover metrics, the analytics report shows a 0.5% failover rate, yet the SIP logs indicate a 4.2% failure rate during peak hours in Singapore. The discrepancy seems to stem from how the analytics engine categorizes 408 timeouts versus 503 errors. We suspect the API is dropping events that occur during the initial SIP registration retry window. The environment is Genesys Cloud Enterprise, and we have verified that the trunk credentials are valid. We need to ensure our reporting accurately reflects carrier performance for SLA compliance. Is there a known issue with the analytics aggregation logic for BYOC trunks in this region, or is this a configuration error on our end regarding the outbound routing rules?