Running Edge 2024.6.1 on Dell R750 nodes here in Tokyo. Primary WAN link flapped again during nightly maintenance window. Cluster automatically flipped to local survivability mode. Local SIP desk phones stayed registered. Predictive routing dialer campaigns went completely dark. Agent status pages show a wall of UNREGISTERED states within 12 seconds of the failover trigger.
Checked the Edge appliance logs. SIP registration retries keep hitting 408 Request Timeout against the local media server. Network team confirmed the bond0 interface stays green. BIOS C-states remain disabled so the CPU isn’t throttling. Predictive routing relies on the primary cloud connection for campaign pacing, but the local routing table should handle the outbound SIP trunks.
API call to GET /api/v2/outbound/campaigns returns a 503 Service Unavailable when the Edge is in survival mode. Response payload shows {“errors”:[{“message”:“Predictor service unreachable”}]}.
Don’t know if it’s a routing table issue or a predictor service timeout. The outbound queue just sits there doing jack all.