Struggling to figure out why the utilization metrics reported in the Genesys Cloud Analytics dashboard diverge significantly from the raw SIP Call Detail Records (CDRs) exported from our carrier gateways for the Singapore BYOC trunks.
We are managing 15 BYOC trunks across the APAC region, primarily Singapore. The environment is running Genesys Cloud v2024.2. The issue manifests specifically during peak hours (10:00 AM - 2:00 PM SGT). The Analytics dashboard reports a peak concurrent call count of approximately 850 sessions on the primary trunk group. However, when we correlate this with the SIP CDRs from our carrier (which logs every INVITE and 200 OK), the actual concurrent sessions recorded are closer to 680. This 20% discrepancy is causing significant issues with our capacity planning and carrier billing reconciliation.
We have verified that the SIP registration status is stable across all 15 trunks. There are no 408 Request Timeout or 407 Proxy Authentication Required errors logged in the Trunk Management console. The outbound routing logic is straightforward, using a simple weighted round-robin strategy without complex failover conditions that might drop calls silently before analytics capture.
We suspect this might be related to how Genesys handles session teardown for calls that are transferred out of the initial trunk group or calls that encounter early media negotiation issues. The documentation on BYOC Analytics Data Latency mentions a potential 15-minute delay in metric aggregation, but this discrepancy persists even after waiting 24 hours for full data reconciliation.
Has anyone encountered similar variances between dashboard metrics and raw SIP logs for BYOC implementations? Are there specific debug logs or API endpoints in the Analytics module that can provide a more granular view of session lifecycle events to pinpoint where these calls are being dropped from the analytics count? We need to ensure our reporting is accurate for our SLA commitments.