Ran into a weird issue today with our IVR routing logic that seems tightly coupled to our weekly Workforce Management schedule push. We operate out of America/Chicago, and every Friday at 10:00 AM CST, our WFM team publishes the new shifts. Immediately following this event, our main Architect flow starts dropping calls during the initial skill-based routing step.
The specific issue is that agents who have just had their shifts updated via the WFM self-service portal are not appearing as available in the Architect queue, even though their status shows ‘Available’ in the Admin console. The flow hits a ‘User Not Found’ error when attempting to retrieve the queue depth for the ‘Support-Tier1’ group. This is incredibly frustrating because the agents are technically ready to take calls, but the routing engine seems to be caching old state from before the publish.
We have verified the following:
- The API call to
/api/v2/architect/flows/{flowId}returns the correct configuration, and the skill groups are properly mapped to the queue. - We manually refreshed the agent profiles in the Admin UI, and the skills updated correctly, but the Architect flow still fails to route to them until we restart the flow or wait 15 minutes for the cache to expire.
The error log in Architect shows a 404 Not Found for the queue entity, which is bizarre because the queue exists and has capacity. This only happens during the 10:00 AM CST publish window. Outside of this window, the flow works perfectly. We are using Genesys Cloud SDK v2.0.15 for our custom integrations, but this seems to be a platform-level routing issue.
Is there a known latency between WFM schedule publishing and Architect queue availability updates? We need a way to synchronize these events or at least introduce a delay in our IVR logic to prevent dropped calls during this critical window. Any insights on how to handle this race condition would be greatly appreciated.