GitHub availability report: June 2026

2026-07-08 · Source: The GitHub Blog · Field: Technology & Digital — Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, medium

Summary

GitHub's June 2026 availability report details infrastructure investments and service performance, including six incidents. Monolith traffic in Azure reached 45% in Central US, below target due to a stability incident pause, with the ramp restarting June 17. Git in Azure increased from 30% to 43% (HTTP and SSH combined), missing its 50% June target, and is expected to plateau near 45% to avoid user latency. Significant progress was made on extracted services: "pullsd" now handles 100% of anonymous pull request reads, "reposd" served 50% of production REST traffic from Azure, and a new users service offloads approximately 500,000 queries per second. API rate limiting is 97% handled at the Gateway, and client-side database load shedding is active on 5% of production traffic. Incidents included Copilot code review failures (June 4), HTTP 504 errors for signed-out users (June 8), GitHub API authentication failures (June 10), Opus 4.8 model degradation (June 16), Copilot frontier chat model unavailability (June 17), and background job service delays (June 25).

Key takeaway

For DevOps Engineers managing large-scale cloud migrations and critical service reliability, GitHub's experience highlights the importance of prioritizing availability over aggressive targets. You should implement strict stability gates for infrastructure ramps, accepting controlled pauses to prevent cascading failures. Additionally, diversify your inference providers and enforce two-person confirmation for production changes to enhance resilience against both internal errors and upstream outages.

Key insights

GitHub prioritizes availability and stability through controlled infrastructure migration and incident learning, accepting target adjustments for reliability.

Principles

• Availability, then capacity, then features.
• Controlled pauses prevent relearning lessons at higher load.
• Reduce reliance on any single inference provider.

Method

Implement stability gates for infrastructure ramps, pin dependency versions, and gradually roll out configuration changes with strong validations.

In practice

• Require two-person confirmation for production access.
• Implement client-side database load shedding.
• Balance capacity across multiple inference providers.

Topics

• GitHub Availability
• Azure Cloud Migration
• Service Reliability
• Incident Management
• Copilot
• Microservices Architecture

Best for: CTO, Director of AI/ML, AI Architect, DevOps Engineer, Software Engineer, VP of Engineering/Data

Related on AIssential

• GitHub availability report: May 2026 — GitHub prioritizes availability and capacity by re-architecting its infrastructure to support rapid AI-driven traffic growth.
• The Pulse: AI load breaks GitHub – why not other vendors? — GitHub's severe reliability issues stem from under-preparedness for AI-driven load spikes and complex infrastructure migration.
• CI/CD with Robert Erez — Effective software delivery balances automation, controlled progressive releases, and adaptable infrastructure, prioritizing practical outcomes over dogmatic approaches.
• GitHub Uses eBPF to Eliminate Deployment Risks and Prevent Circular Failures — eBPF enables kernel-level network control to prevent circular deployment dependencies and enhance system resilience.
• Reliable LLM Inference at Scale — Reliable LLM inference at scale requires sophisticated capacity management, cost-based routing, and robust runtime failure detection.
• The Pulse: is GitHub still best for AI-native development? — GitHub's declining reliability and strategic drift challenge its status as a leading AI-native development platform.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The GitHub Blog.