Pulling queue metrics via /api/v2/analytics/queues/summary with interval=PT1H. The response gives handled, abandoned, and serviceLevel counts, but the math doesn’t match the dashboard. I’m summing handled and dividing by total offered, which seems wrong. How do you calculate the actual percentage from these raw interval buckets? The serviceLevel field just returns a count of calls answered within threshold, not a rate. Trying to replicate the dashboard logic in Python.