Architect v2024.3.1 is routing the inbound flow through a custom data action node to pull active trunk credentials from our internal routing table. Everything works fine on the primary pool. Once the Ohio BYOC failover triggers around 2 PM ET, the data action node throws a DATA_ACTION_EXECUTION_FAILED: CARRIER_AUTH_TIMEOUT. The endpoint being hit is https://api.internal-routing.net/v2/trunk-creds/{region}. Console traces show the request hangs for exactly 4.2 seconds before the proxy drops it. The whole thing is doing jack all during peak Eastern hours. SIP registration on the Ohio pool stays green, but the outbound routing table isn’t updating the carrier weights. The environment runs 15 BYOC trunks split across three regions, and the failover logic swaps the Ohio pool in without a hitch. The issue just sits in that data action step. The flow node retries twice, then pushes the call to the fallback queue. Carrier-specific quirks on the Ohio side seem to be blocking the credential handshake. The payload response comes back with a 401 Unauthorized even though the base64 auth header matches exactly what works on the primary. Network traces show the TCP handshake completes, but TLS renegotiation stalls on port 443. Architect’s built-in trunk health API returns STATUS_DEGRADED right after the failover event. Data action timeouts are messing up the whole routing decision. The custom script can’t even reach the next step. Logs attached below show the exact sequence.
{"node_id": "da_trunk_lookup", "status": "failed", "error_code": "CARRIER_AUTH_TIMEOUT", "retry_count": 2, "timestamp": "2024-05-14T19:42:11Z"}