The outbound flow is choking on the Queue block the second the predictive dialer hits the 1.6x pacing threshold. Org is on the 2024-09-17 release, standard US-East tenant. The setup uses a progressive-to-predictive hybrid for a mid-market compliance push. Pacing looks fine in the campaign admin UI, but the Architect flow logs show a steady stream of 422 Unprocessable Entity responses coming back from the routing engine. Abandonment just jumped to 21% in the last two hours. Queue depth is flatlining at zero. Agents are doing jack all.
The contact list syncs cleanly. DNC flags are updating, callable windows are set to 08:00-20:00 EST. Nothing weird there. Problem shows up when the predictive multiplier crosses 1.5x. The Decision block checks the DNC_STATUS attribute, routes to the Queue block, and suddenly the flow throws the 422 error. Call attempts pile up in the Failed state. Agent utilization drops to 12%.
Resetting the flow version did nothing. Bumping the retry timeout to 8 seconds just delays the crash. Stripped down the IVR menu to a single Queue block. Same result. The API trace points to POST /api/v2/architect/flows/{flowId}/versions/{versionId} returning a validation mismatch on the routingAddress payload. Looks like the predictive dialer is pushing a malformed queue reference when the pacing algorithm recalculates mid-cycle. You’ll see the routing engine reject the payload before it even hits the agent group. It’s throwing the error consistently at 1.58x. Don’t see it in the campaign dashboard until the queue depth hits zero.
Flow ID: 8f3a2c1d-9e4b-4120-aa76-55c89d21f003. Version 14. Queue block UUID: 77b2e1a9-4412-88f3-c901-1123456789ab.
{
"timestamp": "2024-09-18T14:32:11.000Z",
"event": "QUEUE_ROUTING_FAILED",
"error": "422 Unprocessable Entity",
"details": "routingAddress mismatch on predictive dialer payload"
}
Timestamps are completely off by 45 minutes. Flow just stalls out.