Websites and services across the United States and around the world struggled to recover Monday after a problem at Amazon Web Services (AWS) left leading games, publishers, streaming platforms and other apps unusable for millions of people.
AWS, a cloud hosting service that underpins much of the web and everyday online tools, went offline after 3 a.m. because of a problem with one of its core database products, the company said.
The issues were resolved by Monday evening, according to the company.
The outage affected users of sites ranging from Snapchat to the McDonald’s app and Amazon’s Ring doorbell cameras to the gaming platforms Roblox and Fortnite. It underlines the fragility of companies — including financial services — that use cloud-based servers to host their data and how suddenly an unplanned outage can affect businesses around the globe.
AWS first reported a problem at 3:11 a.m. ET and said it was dealing with an “operational issue” affecting 14 different services in its center in northern Virginia, at its U.S.-East-1 Region center.
The Downdetector website received 6.5 million reports that more than 1,000 sites and services around the world were offline.
Snapchat said some users were “having issues,” with some posting on social media that their friends’ lists and daily streaks had disappeared.
United Airlines, T-Mobile, Starbucks and McDonald’s also reported having issues. Delta said it had minor delays after the outage. Medicare also seemed to be affected. Users trying to participate in its open enrollment period said they couldn’t log in to the website Monday afternoon. The British government was also affected, with its HM Revenue and Customs website not accessible to some users. United Healthcare’s provider search tool continued to malfunction throughout Monday.
At 6:35 a.m. ET, AWS said the database problem that caused the outage was “fully mitigated” but warned that there may still be delays.
But at 10:14 a.m. ET, AWS confirmed “significant API errors and connectivity issues across multiple services in the US-EAST-1 Region,” saying it was investigating.
At 4:03 p.m. ET, AWS said that it was seeing improvements in its services and that it was reducing limits it had placed on customer requests, though issues appeared to persist for service like Apple Music and FanDuel into the early evening, according to Downdetector.
The problem stemmed from an error with Amazon’s EC2 internal network, which affected DynamoDB, SQS, Amazon Connect and AWS services, the company said. “The root cause is an underlying internal subsystem responsible for monitoring the health of our network load balancers,” the company said in a statement at 11:43 a.m. ET, referring to technology that distributes traffic across various servers. AWS said it limited new activity requests from customers as it worked to restore full functionality.
The service also went down in 2023 and in 2021, when customers found that they couldn’t access airline reservations and payment apps during the five-hour outage.
This time, social media posts reported problems with cloud-based games, social networks and Amazon’s own services, like Prime Video and Kindle.
Messaging services such as Signal were also affected, prompting Article 19, a nongovernmental organization that promotes freedom of expression, to warn of the democratic consequences of such outages.
“These disruptions are not just technical issues, they’re democratic failures. When a single provider goes dark, critical services go offline with it — media outlets become inaccessible, secure communication apps like Signal stop functioning, and the infrastructure that serves our digital society crumbles,” said Corinne Cath-Speth, Article 19’s head of digital.
Tech mogul Elon Musk also weighed in on the outage’s impact on Signal, calling it “pretty weird” and saying “AWS is in the loop and can take out Signal at any time.”
Coinbase, the largest U.S. cryptocurrency exchange, said its services were unavailable because of the outage.
The trading app Robinhood, the payments app Venmo and the language learning app Duolingo also appeared to be experiencing issues Monday.
Media organizations, including Disney, The New York Times and The Wall Street Journal, were also apparently affected, according to Downdetector. The Associated Press said that it was also experiencing issues and that it had enabled its “AP Backup” system for people to access breaking news.
AWS customers were unable to report the problem because its automated support ticking system was also offline.
AWS stores customers’ data and offers a range of online activities. It is a major player in cloud computing, having recently invested in artificial intelligence products. It’s also highly lucrative — it made $107 billion in the 2024 financial year, 17% of Amazon’s total revenue.