What You Need
- A web browser (Chrome, Firefox, Safari, or Edge)
- Access to GitHub's Status Page
- Basic understanding of service health metrics (optional but helpful)
Step 1: Understand the New Three-Tier Severity System
GitHub recently introduced a Degraded Performance state, joining Partial Outage and Major Outage to form a three-tier classification. This change provides more accurate incident descriptions. Here's what each means:
- Degraded Performance – The service works but may have higher latency, reduced functionality, or intermittent errors affecting a small percentage of requests. It does not count as downtime in uptime calculations.
- Partial Outage – A significant portion of the service is unavailable or severely impacted for many users. This counts as 30% of the incident duration in downtime calculations.
- Major Outage – The service is broadly unavailable affecting most or all users. This counts as 100% downtime.
Previously, even minor glitches were labeled as partial outages, causing confusion. Now you can quickly gauge the real impact.
Step 2: Locate Per-Service Uptime Percentages
On the status page, scroll down to find the list of GitHub services (e.g., Actions, Copilot, Issues). Each service now shows its uptime percentage over the last 90 days. This data is updated regularly and gives a clear snapshot of recent reliability. Look for green bars or percentages; they indicate high availability.
Step 3: Decode How Uptime Is Calculated
The uptime percentage isn't simply raw uptime. Each severity level carries a different weight in the calculation:
- Major Outage – full duration counts as downtime (weight 100%)
- Partial Outage – only 30% of the duration counts as downtime
- Degraded Performance – 0% counts as downtime (it's not considered downtime)
For example, a 1-hour partial outage over 90 days equals 18 minutes of effective downtime in the uptime formula. This method reflects that a partial outage affects fewer users than a major one. You can use this to realistically assess service reliability.
Step 4: Use the Granular Copilot Component
GitHub has added a dedicated component for Copilot AI Model Providers on the status page. This gives specific visibility into availability issues related to the AI models Copilot uses. If you rely on Copilot, check this component for real-time updates on model provider health. It helps you distinguish between a Copilot-specific problem and a general GitHub outage.
Step 5: Respond to Incidents with Better Context
When an incident occurs, note its severity level and the affected service. Use the per-service uptime percentages to see if the incident is an anomaly or part of a pattern. For degraded performance, you can anticipate minor slowdowns without panic. For partial or major outages, check the status page for estimated resolution times and subscribe to notifications for updates.
Tips for Getting the Most Out of the Status Page
- Bookmark the status page so you can quickly access it during incidents.
- Enable notifications via email or webhook to receive alerts about changes in service health.
- Review the 90-day uptime trend before planning major deployments or migrations.
- Use the Copilot component if you rely heavily on AI features – it isolates model provider issues from broader GitHub problems.
- Remember that Degraded Performance does not affect uptime percentages, so a service may appear fully available even if you experience some slowness.