GitHub Says August Outage Was Caused by Capacity Failure as Developer Demand Surges

GitHub has attributed its nearly eight-hour outage on August 17 to infrastructure capacity failures, saying a critical component in its Central US data centre failed to scale as traffic reached a new peak.

The outage lasted 7 hours and 47 minutes, disrupting GitHub.com, authentication, APIs, GitHub Actions, pull requests, issues and Copilot for developers and organisations worldwide.

In a post-mortem report, GitHub CTO Vlad Fedorov said the incident was the company’s second major outage in August, following an August 6 failure involving GitHub Actions.

“The resulting capacity pressure spread through our systems, causing authentication failures and disrupting multiple GitHub services,” Fedorov said.

The resulting capacity constraints spread across GitHub’s infrastructure, triggering authentication failures and disrupting multiple services. GitHub said teams had to reroute traffic, isolate affected infrastructure and restore services in stages.

While most services recovered earlier in the day, Copilot took longer to stabilise. According to GitHub, errors in some Copilot services triggered a client-side retry loop that generated additional traffic during recovery. Engineers had to mitigate the behaviour before safely restoring traffic.

GitHub said neither the August 17 incident nor the earlier outage was caused by a code or configuration change. Instead, “both incidents were capacity failures at their core.”

The company acknowledged that rapidly increasing usage had put additional pressure on its infrastructure. Monthly commits have grown from 1.4 billion in April to 2.9 billion, while merged pull requests and new repositories have also increased substantially.

GitHub said it has added more than 3 million CPU cores, 120 petabytes of high-speed storage and additional network capacity as part of its reliability programme. It is also accelerating its migration to Microsoft Azure.

Azure now handles roughly 58% of GitHub’s platform load, up from 12% in May, while supporting about half of all Git operations.

The company is also changing how its services handle retries to prevent similar incidents from cascading across its infrastructure.

“First, we are applying consistent retry limits, retry budgets, and variable timeouts across service-to-service interactions to prevent retry storms and cascading load,” Fedorov said.

GitHub is additionally isolating critical systems, improving observability and alerting, and developing architecture designed to scale read capacity more directly with demand.

Fedorov acknowledged that the outages have exposed weaknesses in GitHub’s existing operational practices.

“We have made progress, but these incidents make clear that we must accelerate this work,” he said.

Interestingly, on the same day, Elon Musk-owned SpaceXAI, which acquired Cursor, announced Origin, an agent-native Git platform designed to compete directly with GitHub.