Genesys Statistics API: Real-time queue observation data returning stale agent counts?

Hey everyone,

I’ve hit a weird snag while trying to build a real-time dashboard for our WFM adherence tracking. We’re monitoring service levels across multiple queues in the US/Central timezone, and I need live data on waiting counts and available agents. I figured the Statistics API would be the way to go, specifically the /api/v2/analytics/queues/realtime/intervals endpoint.

I wrote a quick Python script using the Genesys Cloud SDK to fetch this data every 30 seconds. Here’s the relevant snippet:

from gen_cloud_py import PlatformClientV2

api_instance = PlatformClientV2.AnalyticsApi()
queue_id = "my-queue-id-123"
interval = "PT1M"

try:
 result = api_instance.get_queue_realtime_intervals(
 queue_id=queue_id,
 interval=interval,
 group_by="queue"
 )
 print(result.entities[0].metrics)
except Exception as e:
 print(f"Error: {e}")

The issue is that the agents.available metric seems to lag by about 5-10 minutes compared to what I see in the Genesys Cloud admin console. I checked the interval parameter and tried switching to PT1S for second-by-second granularity, but the numbers still don’t match the live view. The API call itself succeeds with a 200 OK status, and the JSON payload looks well-formed. For example, one response showed:

{
 "agents": {
 "available": 2,
 "busy": 15
 },
 "calls": {
 "waiting": 3
 }
}

But in the UI, I clearly saw 5 agents available. I’ve verified the OAuth token has the analytics:read scope. Is there a known delay with the Statistics API for real-time metrics? Or am I missing a parameter to force a fresh query? I’ve been staring at the docs for hours and can’t find anything about caching behavior for this endpoint. Any help would be appreciated.