Cloudflare outage cuts access to major websites, services restored

NEW YORK,— A widespread disruption at the web services company Cloudflare began to ease on Tuesday after cutting access to major online platforms for large numbers of users across several regions.

The interruption made it difficult for people to reach X, ChatGPT, Canva, Grindr and other popular services including Kurdish websites, according to reports logged with outage-tracking sites.

Cloudflare, which manages a network carrying a significant share of global internet traffic, said the issue started with internal service degradation detected at about 6:40 a.m. Eastern Time.

The company reported that it had implemented a fix and continued working to restore full operations. It noted that some customers could still notice problems while systems returned to normal.

The event added to a series of recent setbacks for widely used online infrastructure. Last month, a problem at Amazon’s cloud division caused widespread disruptions and forced many well-known sites and apps offline.

That incident left users unable to reach platforms such as Snapchat and Reddit for extended periods, prompting renewed concerns about the vulnerability of large technology providers.

Cloudflare’s stock was down roughly 5 percent in early premarket activity as investors reacted to reports of the failure.

The company operates one of the largest global networks that help websites load more quickly while defending them against cyberattacks and sudden spikes in traffic.

Because of its size and role in keeping internet services stable, interruptions can ripple quickly across many platforms.

Downdetector, which compiles outage data from user submissions and other signals, recorded nearly 5,000 reports at the height of Tuesday’s incident.

Those reports fell to roughly 600 by 8 a.m. Eastern Time, indicating that service conditions were improving. The actual number of affected users remains unclear because the site’s data depends on voluntary reporting.

In a statement, Cloudflare said it detected an unusual surge in traffic to one of its services at 11:20 UTC.

The company explained that the irregular activity led to errors for some traffic passing through its network. Cloudflare said technical teams were fully engaged in restoring normal operations.

(With files from Reuters)