How do I correctly to prevent data gaps in the Genesys Cloud real-time analytics dashboard when a BYOC trunk fails over to a secondary carrier in the Asia/Singapore region? The latency spike causes the call_duration metrics to drop out of the active session window for 15 seconds, breaking our SLA calculations. See [Genesys Docs](https://developers.genesys.cloud/docs/real-time-analytics/byoc-failover) for the current event stream definition.
Cause: The real-time analytics engine resets the session window when the SIP trunk ID changes during failover. In Zendesk, call records were often stitched together post-factum, but Genesys Cloud treats the failover as a new interaction unless explicitly bridged. This creates the 15-second gap where metrics disappear.
Solution: Enable preserveSessionOnTrunkFailover in your BYOC trunk configuration. This tells the platform to keep the original interaction ID active during the carrier switch.
{
"trunkId": "your-byoc-trunk-id",
"settings": {
"failover": {
"enabled": true,
"preserveSessionOnTrunkFailover": true
}
}
}
Also, check your SLA calculation logic. If you are using the call_duration metric directly, consider switching to total_handle_time which is more resilient to these mid-call topology changes. This approach mirrors how we handled ticket updates during Zendesk API outages-by ensuring the core record persists despite backend shifts.