GitHub Copilot Hit 100% Error Rate During January Outage Tied to OpenAI GPT-4.1

GitHub’s January 2026 availability report, published February 11, details two significant incidents that disrupted services for developers worldwide—including a Copilot outage where error rates briefly hit 100% due to complications with OpenAI’s GPT-4.1 model.

The first incident struck on January 13 between 09:25 and 10:11 UTC. GitHub Copilot’s error rates averaged 18% during the 46-minute window, spiking to complete failure at points. Chat features across VS Code, JetBrains IDEs, and other integrated products went down.

A configuration error during a model update triggered the initial problem. GitHub rolled back the change, but recovery stretched until 10:46 UTC because OpenAI was simultaneously experiencing degraded availability for GPT-4.1. The dependency chain failure highlights ongoing reliability concerns as AI coding assistants become embedded in developer workflows.

Two days later, on January 15, GitHub faced broader infrastructure troubles. Between 16:40 and 18:20 UTC, users encountered increased latency and timeouts across issues, pull requests, notifications, Actions, repositories, API access, and account login. Combined web and API request failures averaged 1.8%, briefly peaking at 10% early in the incident.

The culprit? An infrastructure update to GitHub’s data stores. Upgrading to a new major version created unexpected resource contention under production load—the kind of problem that doesn’t surface in testing environments. Unauthenticated users bore the brunt of the impact, though authenticated sessions weren’t spared. Rolling back to the previous stable version resolved the issue.

January proved rough for GitHub beyond these two headline incidents. Authentication services hiccuped on January 22 for nearly an hour. Repository creation and cloning degraded between January 24-25 due to database latency. Actions workflows ran an average 49 seconds late on January 28.

Third-party platforms felt the ripple effects too. Railway reported intermittent GitHub authentication failures and “Repo Not Found” errors from January 26-29 after hitting GitHub’s OAuth token rate limits.

GitHub says it’s implementing stronger monitoring, improved test environments, and tighter configuration safeguards following the January incidents. The company also flagged that February 9 incidents will appear in next month’s report—early February already saw AWS-related issues impacting the platform.

For teams relying on GitHub Copilot as core infrastructure, the OpenAI dependency adds a variable outside GitHub’s direct control. When your AI assistant’s upstream provider stumbles, your developers wait.

Image source: Shutterstock