What this indicator measures
This indicator tracks publicly reported service incidents over a configured observation window.
Calculation method
The counter reads public status information and counts incidents matching the configured period and criteria.
Simplified formula
Incident count = public incidents detected during the configured time windowData sources
Sources include public OpenAI status pages or related service status feeds.
Known limitations
Only publicly reported incidents are included. Minor, regional or internal issues may not appear in the public status feed.
Refresh and interpolation
The incident count is refreshed periodically and should not require interpolation unless displayed as a live rate.