Primary WAN link drops again during the Tokyo ISP maintenance window and the Edge 2024.6.1 cluster on the R750s flips straight into local survivability mode. Platform API calls to GET /api/v2/edges/{edgeId}/status start throwing 401 Unauthorized the second the switchover finishes. Tokens can’t validate against the local proxy even though POST /api/v2/platform/oauth/token returns a clean 200. Local SIP desk phones keep registering without a hitch, so the cluster itself is healthy. The API gateway seems to be ignoring the routing table updates that happen when the WAN flips. Kept the night shift on hold for two hours while tickets piled up.
Curling the endpoint from the R750 management console gives the exact same 401 response. The X-Request-ID header shows the call hits the local edge proxy instead of routing out through the secondary ISP link. Bounced the edge-routing service and cleared the DNS cache on the appliance, did jack all. The health dashboard on port 8090 shows green across the board while the API keeps rejecting valid tokens. Raw response dumps this: {“message”:“Token validation failed: issuer mismatch”,“code”:“UNAUTHORIZED”,“status”:401}. The issuer string in the JWT still points to the primary cloud region even after the local failover completes.