GitHub Overhauls Standing Web page With New Degraded Efficiency Tier

GitHub rolled out vital modifications to its standing web page on April 17, 2026, introducing a brand new incident classification system and per-service uptime metrics that give builders extra granular visibility into platform well being.

The replace comes after the corporate acknowledged availability points earlier this 12 months, and represents a direct response to developer complaints about unclear incident reporting.

Three-Tier Incident Classification

Probably the most notable change is a brand new “Degraded Efficiency” severity degree sitting beneath Partial Outage and Main Outage. Beforehand, even minor service hiccups triggered a Partial Outage classification—deceptive customers into pondering providers have been down once they have been merely gradual.

Beneath the brand new system, Degraded Efficiency covers conditions the place providers stay operational however impaired—assume elevated latency, lowered performance, or intermittent errors hitting a small share of requests. Partial Outage now signifies vital unavailability for a significant variety of customers, whereas Main Outage indicators broad unavailability affecting most or all customers.

Uptime Metrics Get Particular

GitHub now publishes 90-day uptime percentages for every particular person service on its standing web page. The calculation weights incidents by severity: Main Outages rely at 100% downtime, Partial Outages at 30%, and Degraded Efficiency at 0%.

A sensible instance: a 1-hour Partial Outage interprets to 18 minutes of efficient downtime within the calculation, not the total hour. Degraded Efficiency incidents do not affect uptime percentages in any respect—the service technically stayed purposeful.

Copilot Will get Its Personal Incident Monitoring

A brand new “Copilot AI Mannequin Suppliers” element now separates model-specific points from broader Copilot outages. Beforehand, when a single AI mannequin supplier went down, GitHub declared an incident in opposition to all of Copilot—even when options like Chat and cloud agent may fall again to different fashions.

For groups utilizing auto mannequin choice, this distinction issues. A single mannequin going offline would not essentially imply Copilot stops working.

Why This Issues for Dev Groups

For engineering organizations operating CI/CD pipelines by GitHub Actions or relying on Copilot for day by day coding workflows, clearer incident classification helps with capability planning and incident response selections. Figuring out whether or not GitHub is experiencing degraded efficiency versus an precise outage impacts whether or not you wait it out or change to backup processes.

The 90-day uptime metrics additionally give procurement and safety groups exhausting numbers for vendor assessments—one thing GitHub beforehand saved nearer to the vest.

Picture supply: Shutterstock