Metrics
Cards marked Soon are placeholders — the slot is reserved in the nav but the underlying feature isn't built yet.
Cumulative fleet-wide growth: Life-Arc node creation (people, places, events, moments) + Master History-Arc engagement, both by week and by type. Was on the User Account Manager page; promoted here because it's system-wide telemetry.
GET /dev/usage-report ↗ Per-User AI Costs LivePer-user AI spend for the last 30 days — total USD, token counts, and model breakdown. Required before billing goes live. Append ?window_days=N to change the window.
Fleet-wide AI spend trend (daily bars), per-model + per-event breakdowns, and top 5 spenders. Pulls from ai_metrics. Hosting spend (Container Apps, Postgres, Blob) is a separate Soon — needs Azure Cost Management REST.
GET /dev/system-costs ↗ Per-User Usage & Performance LivePer-user activity over a rolling window: AI calls, error rate, p50/p95 latency, sessions, dialog turns, photos, cost. Sortable by any column. 24h / 7d / 30d / 90d / 365d windows.
GET /dev/user-usage