GitHub availability report: May 2026

2026-06-11 · Source: The GitHub Blog · Field: Technology & Digital — Cloud Computing & IT Infrastructure, Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Advanced, long

Summary

GitHub's May 2026 availability report details nine incidents and ongoing infrastructure improvements aimed at enhancing resilience amidst rapid traffic growth, largely driven by AI-assisted workflows. The company is migrating 40% of its monolith traffic to Azure, up from 8% in February, with Git traffic at 30% and repository replication at 99%, effectively doubling capacity in four months. Key structural changes include isolating the primary database cluster, fully cutting over the new users service, and rolling out stateless authentication tokens. Incidents in May included a database schema migration causing 1.3% 5xx errors on May 4, GitHub Actions hosted runner degradations on May 5 and 6 affecting 13.5% of jobs, and a 32-bit integer key overflow halting new pull request review threads on May 6. Other issues involved Copilot session failures, Actions workflow disruptions from a service account suspension, and Copilot degradation due to an upstream API problem affecting GPT-5.2, GPT-5.3-Codex, GPT-5.4, and GPT-5.5 models.

Key takeaway

For AI Architects and MLOps Engineers scaling critical services, GitHub's May 2026 report underscores the necessity of proactive infrastructure resilience. You should prioritize service isolation, robust database management, and comprehensive monitoring to prevent cascading failures. Implement dynamic throttling for migrations and establish allowlists for service accounts to mitigate common incident types. Review your dependency resilience and automated failover strategies, especially for AI models, to ensure continuous availability amidst rapid growth.

Key insights

GitHub prioritizes availability and capacity by re-architecting its infrastructure to support rapid AI-driven traffic growth.

Principles

• Prioritize availability, then capacity, then features.
• Isolate services to prevent cascading failures.

Method

Transform infrastructure by migrating to Azure, breaking monoliths into isolated services, and eliminating shared failure points, including database cluster isolation and stateless authentication tokens.

In practice

• Implement dynamic throttling and circuit breakers for database migrations.
• Add allowlists for critical service accounts to prevent suspension.
• Strengthen monitoring for column size limits and replication lag.

Topics

• GitHub Availability
• Infrastructure Resilience
• Azure Migration
• AI Development Workflows
• Database Management
• GitHub Actions
• GitHub Copilot

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

Related on AIssential

• 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.
• How I Built a Commit-Aware Multi-Cloud Reliability Engine (And What Two Patents Taught Me) — Code commits provide early signals for predicting cloud failures and enabling proactive reliability measures.
• GitHub Acknowledges Recent Outages, Cites Scaling Challenges and Architectural Weaknesses — GitHub publicly addressed a series of recent availability and performance issues in early 2026, attributing them to rapid growth, architectural coupling, scaling limitations, and an inability to effectively shed load.
• AI-ready infrastructure: The foundation for scalable AI workloads — AI success depends on robust infrastructure capable of rapid data movement, predictable scaling, and reliable operations, rather than solely on model strength.
• Claude outage, June 2026: Reckoning with AI’s increasing status as infrastructure — Generative AI is critical infrastructure; single-vendor dependency poses a significant business continuity risk.
• How the lakebase architecture stays resilient to cloud failures — Lakebase achieves high availability for agentic workloads through stateless compute, zone-redundant storage, and a refactored control plane.

Open in AIssential →

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