SIP Trunk Registration Drops During WFM Weekly Schedule Publish

Just noticed that our SIP trunk registrations are dropping precisely when the weekly WFM schedule publish job completes in Genesys Cloud 2024-2.0. This is causing a brief window of unreachability for agents who are scheduled to be online but whose devices have reset their registration state.

  • The issue occurs every Friday at 11:59 PM Central Time when the automated publish workflow triggers.
  • Agents using the Genesys Cloud Desktop application (version 7.2.1) report a “Registration Failed” error immediately after the schedule goes live.
  • Our SIP trunk provider logs show a burst of 408 Request Timeout responses from the Genesys Cloud edge nodes during the 5-minute publish window.
  • The GET /api/v2/wfm/schedules/{scheduleId}/history endpoint confirms the publish was successful, but the telephony layer seems to be resetting session states.
  • We have verified that agent availability is correctly updated in the WFM module, so this is not a data sync issue between WFM and the telephony provisioning engine.
  • The problem is specific to the weekly publish; daily ad-hoc schedule changes do not trigger this SIP re-registration storm.
  • We are using the default SIP trunk configuration provided by Genesys Cloud, with no custom codec restrictions or advanced security profiles applied.
  • The impact is minimal but annoying; agents have to manually refresh their softphone connections, which disrupts their start-of-week workflow.
  • We have checked the system health dashboard, and all telephony services are reporting green status during the incident window.
  • This behavior started appearing after the last platform update, which included changes to the WFM schedule optimization engine.
  • We need to understand if there is a known race condition between the WFM publish API and the SIP registration service.
  • Is there a way to stagger the SIP re-registration requests to avoid this timeout burst?
  • We are considering disabling the automatic publish and using manual overrides, but that defeats the purpose of our automated scheduling workflow.
  • Any insights into how the WFM publish job interacts with the telephony provisioning backend would be appreciated.
  • We are open to testing any configuration changes that might mitigate this SIP trunk instability during high-traffic schedule updates.