Is it possible to accurately derive a real-time Service Level percentage directly from the raw interval data returned by the /api/v2/analytics/conversations/queues/events endpoint? I am building a custom monitoring script that polls this endpoint every sixty seconds to feed threshold checks into my PagerDuty integration. The response returns bucketed metrics like offerCount, answerCount, and answeredWithinSLA, but the documentation does not specify how these aggregate across intervals. When I sum the values manually, the percentage drifts from the dashboard. Here is my current calculation logic: sla_pct = (sum(interval.answeredWithinSLA) / sum(interval.offerCount)) * 100. I also notice the JSON payload includes a total object at the root level, which seems to mirror the summed values. Should I rely on total.answeredWithinSLA / total.offerCount instead of iterating through the intervals array? The [Genesys Analytics API Reference](https://developer.genesys.cloud/apidocs/analytics/v2/events) mentions that interval data is sampled, but does not clarify if SLA thresholds are evaluated per-interval or globally. I need to ensure my auto-escalation webhook triggers only when the actual SLA breaches the 80 percent target over a rolling five-minute window. Currently, my script returns a 200 status but calculates roughly 12 percent higher than the web UI. I suspect the offerCount denominator includes abandoned calls that should be filtered out, yet the payload does not expose an abandonedCount field in the same structure. How should I adjust the division logic or filter the JSON response to match the official SLA calculation? I am using a standard OAuth2 client credentials flow and sending the request with Accept: application/json. Any corrections to the arithmetic or endpoint parameters would help align my integration.