Hourlong disruption affected sites including New York Times, U.K. government portal; cloud-services provider Fastly reports internal problem
Dozens of websites in the U.S. and Europe briefly went dark Tuesday, victims of an internal glitch at a major cloud-service provider and the latest example of how a problem at a single player in the internet’s piping can have outsize repercussions.
Sites including the U.K. government’s main public-services portal and several major U.S. and European news outlets, such as CNN, the New York Times NYT -1.46% and Le Monde, were inaccessible to at least some users for roughly an hour, beginning around 6 a.m. EDT.
Those sites went down after Fastly Inc., the operator of a content delivery network service, said it had experienced a configuration problem that affected its infrastructure globally. Websites use Fastly and others like it to speed up their webpages. Content-delivery networks essentially store content from those sites on a large number of servers that are closer to end users, reducing the time it takes information to reach them. That speeds up things like website loading and streaming.
The setup, though, can make swaths of the internet vulnerable to failures that affect a single delivery network. In a securities filing, Fastly said it had 336 enterprise clients as of March 31. Tuesday’s interruption caused Fastly’s observed traffic volume to drop by about 75%, according to network monitoring company Kentik Inc.
“The incident highlights the reliance of many of the world’s biggest websites on content delivery networks such as Fastly,” said Toby Stephenson, chief technology officer of British IT services company Neuways Ltd.
Shortly before 7 a.m., Fastly said it had applied a fix to the problem, and websites began to come back online. Less than an hour later, it reported all services had been restored and called the incident resolved.
Fastly Chief Executive Joshua Bixby, reached in the middle of the outage, said the problem wasn’t a cyberattack.
The outage is the latest in a series of interruptions to the availability of popular online services. As more businesses outsource digital infrastructure to outside providers, a technical failure at one of them—which can include giant companies like Alphabet Inc.’s Google or smaller, less well-known companies like Fastly—can ripple across the internet.
In December, a login glitch made more than a dozen Google services such as Gmail and Google Docs inaccessible for roughly an hour, disrupting workplaces and schools.
A crude 2016 cyberattack against Dyn, a service that acts as a virtual directory for internet addresses, took down hundreds of websites.
Fastly, founded in 2011, is one of a number of prominent content-delivery network companies, competing with companies like Akamai Technologies Inc. and Cloudflare Inc. Fastly went public in 2019 and has reported fast-growing revenue, rising to $291 million in 2020, though it is still loss-making.
The outage occurred well before the stock market opened. With most affected sites back up by the opening bell, Fastly shares jumped 11% Tuesday.
Among Fastly clients that experienced at least some outages was the U.K.-based food-delivery service Deliveroo PLC. One Twitter user complained about waiting for nearly an hour and a half for food without any response from the company.
A Deliveroo representative apologized via the company’s Twitter account, writing, “As our systems are down due to a widespread internet outage, we can’t process refunds or check on the order status.” The company didn’t immediately respond to a request for comment.