Methodology · v2.6

SRIService Reliability

Published service-level uptime percentage for {legal_name}'s primary service(s). EXTRACTION RULES — MANDATORY SEARCH ORDER (do not bail until step 4 fails): (1) Try the company-owned status page using ALL these canonical URL patterns: `<company-domain>/support/systemstatus`, `<company-domain>/status`, `status.<company-domain>`, `statuspage.io/<company-slug>`, `instatus.com/<company-slug>`. Most service / platform / hybrid companies of any size maintain one of these — a site-restricted web search (site:status.<domain>) is required if the direct URL 404s. (2) If no status page, search for SLA-declared uptime guarantee in the company's Terms of Service / Master Services Agreement / Service Level Agreement PDF on its own domain. (3) If no SLA, query Downdetector at downdetector.com/status/<company-slug>/ and convert reported incidents in last 12 months to uptime proxy: 0-10 incidents → 99.95, 11-50 → 99.9, 51-200 → 99.5, 200+ → 98.5 (state proxy method in _reasoning). (4) If no Downdetector page, query UptimeRobot / Pingdom public history of the primary domain. Prefer historical-observed over SLA-declared when both exist. ONLY return type="absent" with reason="not_disclosed" after demonstrating in _reasoning that steps 1-4 ALL returned no data — explicitly list the URLs tried. Returning absent without listing tried URLs is a protocol violation. Return exact_numerical{value: <uptime_pct_0_100>, unit: "percent", source_urls: [...]}. Server applies ratio_band (<99 → 30, 99-99.9 → 60, 99.9-99.99 → 80, 99.99+ → 95).