No idea why this is happening, the Workforce Management API is rejecting our adherence data payloads during a BYOC trunk failover simulation. The system operates 15 BYOC trunks across the Asia/Singapore region, and while voice routing handles the carrier switch seamlessly, the analytics pipeline breaks.
When the primary trunk drops and traffic shifts to the secondary carrier, the WFM API endpoint /api/v2/wfm/schedules/adherence starts returning a 400 Bad Request error. The error message indicates a malformed timestamp format, yet the payload structure remains identical to the primary trunk scenarios. The SDK version is 2.4.1, and we have verified that the NTP sync on the SIP proxies is accurate to the millisecond.
The issue appears strictly tied to the carrier-specific quirks of the secondary provider, which introduces a slight delay in the SIP 200 OK response. This delay seems to desynchronize the agent state updates in Architect, causing the WFM service to receive overlapping or out-of-order events. We have checked the outbound routing logic and confirmed that the failover logic is correct. Is there a known limitation with the WFM API regarding timestamp precision during rapid carrier switches, or should we adjust the retry logic in our integration?