SIP Registration Drops During Peak Shift Swap Validation

Could use a hand troubleshooting this weird interaction between WFM schedule adherence checks and our SIP trunk stability. We are running Genesys Cloud 2024.3 (US-East-1) and use a custom BYOC edge. Whenever the WFM module runs its nightly adherence report and validates pending shift swaps, our SIP registration rate drops by about 15% for roughly 2 minutes.

Here is what we are seeing:

  1. The WFM adherence job triggers at 06:00 AM CT to validate shift swaps for the next business day.
  2. Simultaneously, the /api/v2/wfm/schedules/shiftswaps endpoint starts polling for agent confirmation status.
  3. Within 30 seconds, the SIP trunk logs show a spike in 408 Request Timeout errors for agents logging into their softphones.
  4. The WebRTC handshake fails for approximately 40 agents, forcing them to restart their desktop application.

This only happens when the shift swap queue exceeds 100 pending items. I have checked the network latency to the Chicago region, and it is stable. The SIP trunks are configured with a registration interval of 300 seconds. Could the WFM validation process be consuming too many concurrent API threads, causing a bottleneck in the signaling path? Any insights on how to decouple these processes would be appreciated.