GitHub recently faced two significant outages that fell short of the reliability standards users expect. In response, the platform has been overhauling its infrastructure to handle an explosion in developer activity driven by agentic workflows. This article answers common questions about GitHub's challenges, the steps taken to improve availability, and what users can expect going forward.

What prompted GitHub to significantly increase its capacity plans?

GitHub initially planned to boost its capacity tenfold by October 2025 as part of a broader reliability and failover initiative. However, by February 2026, it became clear that the platform would need to scale to 30 times its current size. The primary catalyst was a sharp acceleration in agentic development workflows starting in the second half of December 2025. Metrics such as repository creation, pull request activity, API usage, automation, and large-repository workloads all surged simultaneously. This rapid shift in how software is built forced GitHub to rethink its growth projections and prioritize scalability far beyond original estimates.

How does exponential growth affect GitHub's systems?

Exponential growth doesn't stress a single component; instead, it creates cascading effects across interconnected services. For example, a single pull request can touch Git storage, mergeability checks, branch protection, GitHub Actions, search, notifications, permissions, webhooks, APIs, background jobs, caches, and databases. At high scale, small inefficiencies compound: queues deepen, cache misses turn into database load, indexes fall behind, retries amplify traffic, and one slow dependency can degrade multiple product experiences. This complexity makes it difficult to predict failure points and requires systematic isolation of critical paths.

What are GitHub's top priorities for reliability?

GitHub has established a clear hierarchy: availability first, then capacity, then new features. This means eliminating unnecessary work, improving caching mechanisms, isolating critical services, removing single points of failure, and migrating performance-sensitive code to systems built for modern workloads. The focus is on reducing hidden coupling between services, limiting blast radius during failures, and ensuring the platform degrades gracefully when one subsystem is under pressure. These principles guide every architectural decision, from database optimization to service decomposition.

What immediate steps did GitHub take to address bottlenecks?

In the short term, GitHub resolved several bottlenecks that appeared faster than anticipated. Key actions included moving webhooks to a different backend to relieve MySQL pressure, redesigning the user session cache, and redoing authentication and authorization flows to substantially reduce database load. Additionally, GitHub leveraged its migration to Azure to provision significantly more compute capacity. These fixes were critical for stabilizing the platform while longer-term solutions were developed.

How is GitHub isolating critical services to reduce failures?

GitHub is systematically isolating critical services like Git and GitHub Actions from other workloads. This work began with a careful analysis of dependencies and traffic tiers to understand what needs to be separated and how to minimize the impact of attacks on legitimate traffic. Each risk is addressed in order of severity. The goal is to minimize blast radius—ensuring that a problem in one subsystem doesn't bring down unrelated services. This approach reduces the likelihood of widespread outages.

What architectural changes is GitHub making for long-term reliability?

Beyond immediate fixes, GitHub is migrating performance- and scale-sensitive code from its Ruby monolith to Go, a language better suited for high-concurrency workloads. The company is also accelerating the move from its smaller custom data centers to public cloud infrastructure, with an eye toward a multi-cloud strategy. These changes reduce dependence on any single provider and improve resilience. By decoupling services and adopting cloud-native patterns, GitHub aims to handle future growth more gracefully.

Why did GitHub upgrade its capacity target from 10X to 30X?

The upgrade from a 10x to a 30x capacity target was driven by data showing that agentic development workflows were accelerating faster than anticipated. By nearly every measure—repository creation, pull request activity, API usage, automation, and large-repository workloads—the growth rate exceeded projections. This exponential increase required a more aggressive scaling plan to ensure reliability. GitHub recognized that designing for a future requiring 30 times today's scale would better prepare the platform for the ongoing transformation in software development.