Returns the uptime rate over a time window and the downtime incidents recorded in it. The rate is computed from confirmed incidents only, never from raw ping samples: a single failed ping can be a hosting provider blocking our monitoring IPs, so it takes a multi-region confirmation before anything counts against uptime. A project whose monitoring is off or paused returns monitoring_enabled false and a null rate rather than a fabricated 100. The window defaults to the last 30 days and cannot exceed 366 days. Incidents are truncated to limit, most recent first; incidents_total reports how many the window holds, and the rate always covers the whole window regardless of that truncation. An incident carries the region that observed it, lowercased (eu_west, us_east, us_west, asia_hk, australia), and description is a machine-readable reason such as dns_resolution_failed. Neither is a closed set: treat an unknown value as opaque.