Stuck on a weird SIP registration issue that seems tied to our weekly WFM publish cycle. We are running Genesys Cloud v24.5.0 in the Chicago environment. Every Tuesday, when we push the updated schedules via the WFM API, agents who have approved shift swaps start experiencing intermittent SIP registration drops. Their status in WFM shows ‘Available’, but the softphone fails to register with the Genesys SIP trunk, throwing a 408 Request Timeout after 30 seconds.
The standard agents without swaps register fine. It feels like the WFM state change isn’t syncing properly with the telephony provisioning layer for swapped shifts. We’ve checked the SIP trunk logs and see the initial INVITE going out, but no 200 OK coming back from the Genesys edge.
Here is the relevant SIP trunk configuration we are using:
trunkConfiguration:
sipTrunkId: "trk_chi_main_001"
protocol: "SIP"
transport: "TLS"
registrarTimeout: 30000
retryInterval: 5000
maxRetries: 3
authMethod: "DIGEST"
userDomain: "genesyscloud.com"
registrarUrl: "sip:registrar.genesyscloud.com:5061"
allowedCodecs:
- "PCMU"
- "PCMA"
- "G729"
wfmSyncEnabled: true
shiftSwapAware: true
The shiftSwapAware flag is enabled, which we thought would handle the dynamic schedule changes, but it seems to be causing a race condition. The agents are stuck in a ‘Registering’ state until they manually refresh their profile or wait for the next WFM sync cycle (which is every 15 minutes). This is killing our adherence scores for the morning shift.
Has anyone else seen this behavior with shiftSwapAware enabled? Should we be using a different approach for handling telephony registration for agents with dynamic schedules? We are considering disabling the flag and forcing a full re-provision, but that seems like a heavy-handed workaround.