Cloudflare Outage History

Cloudflare experienced an issue where customers encountered 5xx errors after completing Pages deployments, affecting Cloudflare Sites and Services and Pages components. The problem was identified and a fix was implemented within 1.8 hours. Customers experiencing the errors were advised to redeploy their pages to resolve the issue.

Cloudflare experienced delays in audit log processing that affected v2 audit logs for 6.9 hours, while v1 audit logs remained unaffected. The issue impacted the timely delivery of audit data to users. Cloudflare identified the root cause, implemented a fix, and monitored the results to resolve the processing delays.

Cloudflare experienced a 59-minute outage affecting their Dashboard and APIs due to an issue with their entitlements service, causing API requests to fail with 503 errors and preventing users from creating or modifying services. The incident was resolved by rolling back the affected service, with recovery being gradual as in-memory caches repopulated. CDN traffic serving, security features, and already-deployed compute/storage services remained unaffected throughout the incident.

Cloudflare One experienced DNS resolution failures for .internal zones when served by private DNS resolvers accessible only through Cloudflare One, affecting WARP and Magic WAN services. The issue was identified and resolved within 15 minutes through an implemented fix that restored normal DNS functionality.

Cloudflare experienced issues with starting new containers that affected their Sites and Services and Containers components for 1.1 hours. The company investigated the problem and worked to understand the full impact while implementing mitigation measures. The incident was classified as minor and has been resolved.

Cloudflare experienced an issue with delayed audit logs for their Access service, affecting Cloudflare Sites and Services and Access components. The incident lasted 2.3 hours before a fix was implemented and monitoring began. The delay in audit logs impacted the timely availability of security and access monitoring data for users.

Cloudflare Workers experienced an issue where subrequests were returning 1006 errors, affecting Cloudflare Sites and Services along with the Workers platform. The problem was identified and resolved with a fix implementation, with the incident lasting 2.4 hours while Cloudflare monitored the results.

Cloudflare experienced increased HTTP 502 errors affecting users in Buenos Aires, Argentina for 18.8 hours. The incident impacted the Latin America & Caribbean region, specifically the Buenos Aires (EZE) data center. A fix was implemented and the service was restored to normal operation with monitoring in place.

Cloudflare experienced a minor service incident affecting their Access AI control functionality, specifically impacting MCP (Model Context Protocol) server creation and authentication through the Zero Trust Dashboard. Customers were unable to add new MCP servers or authenticate existing ones during the 1.7-hour outage. The issue has been resolved after investigation by Cloudflare's engineering team.

Cloudflare experienced intermittent authentication failures for users logging into the Dashboard using SSO with Okta as their identity provider, with affected users encountering "failed to fetch" errors during login. The issue was isolated to Okta SSO authentication only, while other identity providers, direct username/password login, and Cloudflare Access remained unaffected. The incident was resolved after 55 minutes of investigation by Cloudflare Engineering.