GitHub April 2026 Availability Report: 10 Incidents, Key Lessons

TL;DR

• GitHub experienced 10 incidents in April affecting code search, Copilot, Pages, Codespaces, and core services
• Two major outages: April 23 DNS cascade (5–7% traffic) and April 27 search saturation from scraping (65% timeouts)
• Root causes: automation failures, rate limiting bugs, infrastructure changes, and coordinated abuse traffic

What Dropped

GitHub published its April 2026 availability report detailing 10 incidents that degraded service across multiple products. The most significant were a DNS infrastructure failure on April 23 that cascaded across Copilot, Actions, Git Operations, and Webhooks, and a coordinated scraping attack on April 27 that saturated search load balancers and impacted Issues, Pull Requests, and Projects.

The Dev Angle

If you rely on GitHub's core services, this report matters. April 23's DNS incident affected approximately 5–7% of overall traffic, with Copilot seeing ~7% AI model request failures and Actions experiencing workflow status delays up to 8 seconds. April 27's search outage was worse: up to 65% of searches timed out or errored during peak saturation, directly breaking Issues, Pull Requests, Projects, and Dependabot Alerts.

The April 9 Copilot incidents are particularly relevant if you use Copilot Cloud Agent. A rate limiting bug incorrectly applied limits globally instead of per-installation, causing 84% of new agent session requests to be delayed with queue times hitting 54 minutes. The second wave was compounded by a caching bug that persisted the rate-limited state beyond the actual window.

Code scanning, project boards, and Codespaces also went down. April 20's serialization error prevented code scanning default setup and code quality analyses from triggering on new PRs. April 16's Codespaces incident blocked ~40% of VS Code connections (SSH was unaffected). April 13's GitHub Pages saw 17.5 million failed requests when an automated DNS tool deleted a critical record.

Should You Care?

Yes, if you depend on any of these services. GitHub's transparency here is solid—they're publishing detailed root causes and concrete follow-up actions. The April 23 DNS incident reveals a systemic risk: a single infrastructure change cascaded across six different product areas. GitHub is now investing in safer rollouts, dedicated test environments for infrastructure changes, and faster automated detection.

The April 27 scraping attack (600,000+ unique IPs, 30% of daily search traffic in four hours) shows GitHub's public APIs are under sustained abuse pressure. They're adding traffic restrictions for anonymous users and improving connection handling on load balancers. If you're building integrations that hit GitHub's search or API endpoints, expect tighter rate limits and monitoring.

For Copilot users: the April 9 and April 22 incidents highlight that AI agent services are still stabilizing. April 22's Copilot Chat outage lasted 3+ hours due to a database connectivity issue from an infrastructure config change. These are growing pains, but they're worth tracking if you're betting on Copilot for critical workflows.

What's Next

GitHub is making three categories of improvements: infrastructure resilience (DNS failover, zone-tolerant routing, safer automation), faster detection (new monitoring for DNS failures, rate limit exhaustion, serialization errors), and abuse mitigation (anonymous traffic controls, connection optimization). They're also auditing other systems for similar failure modes.

The report links to deeper posts on April 23 and 27 incidents and status page improvements. Watch the GitHub status page for real-time updates.

Source: GitHub Blog