Breaking: GitHub Revamps Status Page to Improve Transparency During Service Disruptions
GitHub announced today a series of major updates to its status page, introducing a new Degraded Performance severity level and publishing per-service uptime statistics for the first time. The changes, effective immediately, aim to give developers clearer, more accurate information about platform health and incident impact.
“Developers rely on GitHub to build and ship code,” said a company spokesperson. “These updates ensure that when something goes wrong, we communicate the exact nature and scope of the issue—no more, no less.” The moves come after earlier promises to improve reliability and incident communication.
New ‘Degraded Performance’ Tier Improves Incident Classification
GitHub’s status page previously offered only two severity levels: Partial Outage and Major Outage. Now a third tier—Degraded Performance—sits between them, capturing situations where a service is operational but impaired.
“In the past, even minor latency or intermittent errors were labeled as a partial outage,” the spokesperson said. “This misled users into thinking the service was down when it was still usable. The new tier fixes that.” Degraded Performance covers elevated latency, reduced functionality, or errors affecting a small percentage of requests—without counting as downtime in uptime calculations.
Per-Service Uptime Metrics Now Public
For the first time, GitHub is publishing 90-day uptime percentages for each individual service directly on the status page. Users can quickly see the recent reliability track record of services like Actions, Copilot, Issues, and more.
The uptime calculation uses weighted downtime: Major Outage counts 100% of duration; Partial Outage counts 30%; and Degraded Performance counts 0%. For example, a 1-hour Partial Outage results in 18 minutes of effective downtime. “This gives a more honest picture of service health,” the spokesperson noted.
Granular Insights for Copilot AI Model Providers
GitHub is also launching a dedicated Copilot AI Model Providers component on the status page. This component will provide clearer communication about availability issues affecting the underlying AI models that power GitHub Copilot.
“Previously, a model provider outage could be conflated with a broader Copilot issue,” the spokesperson explained. “Now we can isolate and communicate the exact provider impact.” This is part of a broader effort to offer more granular insights into service disruptions.
Background: Why These Changes Matter
GitHub has faced scrutiny over the past year due to several high-profile availability incidents. In response, the company invested heavily in infrastructure reliability and pledged to improve incident communication. The old status page classified all disruptions as partial outages, eroding trust.
“We heard loud and clear that developers wanted more nuance,” the spokesperson said. “A simple ‘up or down’ binary wasn’t enough.” The new system aligns with industry best practices for status page transparency.
What This Means for Developers
Developers can now make smarter decisions during incidents: a Degraded Performance label means they can likely continue working, while a Major Outage signals a critical break. The per-service uptime data enables engineering teams to assess GitHub’s reliability for their specific workflows.
Long term, these changes set a higher bar for transparency across the industry. “We’re committed to being clear about our platform’s health,” the spokesperson added. “This is just the beginning of ongoing improvements.”