Predictive Routing Queue Depth vs. ServiceNow Incident Creation Threshold

Can anyone clarify the exact trigger mechanism for queue depth alerts in Predictive Routing when mapped to ServiceNow via Data Actions? The documentation states:

“Alerts fire when active conversations exceed the defined threshold for 60 seconds.”

However, our MID Server receives duplicate payloads every 10 seconds during peak load. Is this a polling interval issue in the webhook configuration, or does GC batch these events differently? We are seeing HTTP 429 errors on the SN side due to the frequency.