Internet Status — February 22, 2026: Recovering from Cloudflare BGP Outage; Minor Issues Persist
Today the internet is largely stable, though one active issue persists and a significant outage from earlier this week has been fully resolved.
Active Issues:
• Cloudflare Newark (EWR) datacenter continues to experience elevated HTTP request latency, ongoing since February 16. Some Data Loss Prevention (DLP) customers are seeing intermittent errors. Cloudflare is investigating with fixes in progress.
Recently Resolved (Past 72 Hours):
• On February 20, Cloudflare experienced a major outage when approximately 1,100 BYOIP (Bring Your Own IP) prefixes were unintentionally withdrawn via BGP due to a configuration change error. The incident lasted 6 hours and 7 minutes, making affected customer services unreachable from the internet. A full postmortem has been published at blog.cloudflare.com. This was not caused by a cyberattack.
• Several related Cloudflare services were also impacted on Feb 20–21, including Workers AI (elevated 429 errors), Bot Management (detection issues), Analytics Engine API (500 errors), and Durable Objects (free tier invocation failures). All have been resolved.
• GitHub Actions experienced extended job start delays for larger hosted runners on February 20, now resolved.
All Clear:
AWS, Google Cloud, Azure, Fastly, Akamai, OpenAI, and GitHub are all reporting normal operations today.
Overall: One localized Cloudflare issue persists, and the effects of Thursday's major BGP outage have been fully resolved. The internet is up.