Get Analytics

Get full analytics data for the analytics page.

Perf note (2026-04-23): the five queries below (monthly counts ×2 + 30-day paginated aggregations ×3 over message_history) used to run serially on the request thread. They are now executed in parallel via asyncio.gather + asyncio.to_thread — each inner function still owns its own try/except so a single sub-query failure degrades to safe defaults without taking the whole endpoint down. Wall-clock improves roughly Σ latencies → max latency for this endpoint, which is the dominant cost for /dashboard/insights.