'We let you down': GitHub pledges to scale up before developers give up

GitHub's executives expressed regret over a recent platform outage that disrupted developers worldwide on August 17. CTO Vladimir Fedorov apologized for the service failure, which lasted 7 hours and 47 minutes, affecting actions, pull requests, issues, Copilot, and APIs. The incident followed another outage on August 6 involving Actions.

GitHub has been struggling with scalability issues, with monthly commits reaching 2.9 billion, 24 million new repositories, and 130 million merged pull requests. Microsoft Azure handles 58 percent of GitHub's platform load and 50 percent of all Git operations. Fedorov revealed that GitHub is accelerating the migration of more workloads to its parent company's cloud, with the next milestone being an architecture that can scale read capacity linearly with the number of readers.

Before this architecture is implemented, GitHub needs to address scaling weaknesses and strengthen system isolation, retry limits, and monitoring for early signs of traffic spikes. The repeated outages have disappointed developers, who rely on GitHub to build, ship, and operate their work. Fedorov emphasized the platform's responsibility to rectify the issues and rebuild trust through improved scalability and reliability.