Problem
The dashboard UI needs live queue observation metrics, so we’re querying for queue observations every thirty seconds. The response keeps returning { "metrics": { "waiting": { "count": 0 } } } even though the supervisor console shows twelve callers queued. We’ve verified the queue ID matches the routing config. The interval parameter seems to be ignored entirely.
The realtime endpoint is notoriously flaky for high-frequency polling. Switch to querying queue observations instead. It’s designed for live UIs and updates way faster.
1 Like