Amazon Web Services provides cloud service support for millions of websites. So when an outage hit more than 100 of its services early Monday morning, it took down a vast portion of the internet with it. The results were disruptive to say the least.
According to online outage tracker DownDetector, WhatsApp, Venmo, Hulu, Coinbase, Roblox, Slack, Duolingo, the Starbucks and McDonald’s apps, and the United Kingdom government’s official website were all down at one point this morning. Zoom shared incident status reports confirming disruptions, and Fortnite said that the outage was disrupting log-ins. Signal chief Meredith Whittaker confirmed on Bluesky that the messaging app was down due to the AWS outage. Amazon’s own services suffered too, including the Ring safety cameras.
According to numerous complaints on social media, the outage also sent airports into a frenzy. The United and Delta apps were down, and users were unable to check in for flights.
A United system outage caused by the AWS crash also reportedly caused long lines and disrupted check-ins and bag drop-offs across airports in the United States, according to the airline’s own X posts in response to customer complaints. The outage is yet to lead to massive delays, but Delta reported facing some minor ones.
The outage comes as airport operations are already strained due to the government shutdown and an ongoing air traffic controller staffing shortage that has been years in the making. Most airport workers are currently working without pay as the government shutdown marks its nineteenth day on Monday.
AWS is still investigating why the outage happened but the core issue seems to be a problem with the Domain Name System (DNS) resolution of its DynamoDB service, a database offering from AWS. The bottom line: Thousands of companies that relied on a database offering of Amazon suddenly started facing trouble translating user-friendly domain names into machine-readable IP addresses.
The issue stemmed from an DynamoDB API endpoint in the US-EAST-1 region, which has its physical infrastructure centered in northern Virginia, and has since affected other services that use the same regional infrastructure.
The outage began at around 3 am ET. As of 10 am ET, the underlying issue was “fully mitigated,” but Amazon said that only 37 out of the 114 service outages caused by that issue had been resolved.
Although still quite disruptive, it’s an arguably more contained outage (at least for now) than last year’s Crowdstrike outage that sent the world reeling after a faulty update. It took several days for all systems to return to normal functioning. That outage caused thousands of flight cancellations and delays, and cost Delta a whopping $500 million.
Outage debacles often underline concerns about an over-reliance on a select few companies to provide the systems that keep the internet running. When a few providers supply the infrastructure of the entire web, even the smallest thing going awry can have a cascade effect on systems and services worldwide, disrupting the everyday lives of millions. It’s something we take for granted, but Monday morning’s outage was yet another reminder of the pitfalls of our current configuration of the internet and its providers.
