As the US East Coast awoke Monday morning, Amazon Web Services employees in Northern Virginia were likely already hard at work, scrambling to address an issue that took down popular apps and internet services across the globe. 

Difficulty at US-East-1, the largest and most active AWS data center cluster, was cited as the reason behind the outage, which impacted dozens of platforms, from ChatGPT to Signal to Coinbase to Fortnite, per The Verge and other reporting.

Services began to recover around 6:30 a.m. ET, after about 3 hours of hard downtime, though the company warned that delays could continue, according to NBC News.

So what exactly is that data center, what happened to it, and why did that cause such widespread issues? Read on.

What is US-East-1?

Every day, people send large files through Gmail, post on Facebook, shop on Amazon, download movies from Netflix and do countless other things that generate data. All that information has to go somewhere, and that place is what we call “the cloud.”

But despite the name, the cloud isn’t a faraway place in the sky. When someone stores data “in the cloud,” it’s still saved on physical disk drives, just not on your personal device. Instead, it’s housed in servers located in data centers — and US-East-1 happens to be one of the biggest. 

US-East-1 dates back to Amazon Web Services’ origins, powering the infrastructure behind the computing giant. Made up of several centers, the Amazon campus runs through Virginia. Built in 2006, US-East-1 powers many of the biggest websites we use because of its established reputation — but years of recent outages are starting to hurt that. 

Why is it located in Northern Virginia?

Virginia has the largest data center market in the world, comprising more than 300 with a massive concentration in Loudoun County’s self-named “Data Center Alley.”

Development has clustered in NoVa in part because of proximity to the federal government. Local tax incentives also play a role: Projects with a minimum new capital investment of $150 million that create 50 new jobs can receive a sales and use tax exemption on equipment and software. In 2021, this shifted to a requirement of only 10 employees and $70 million for data center builds in localities with higher than average unemployment and poverty rates.

Those incentives have made it easy for US-East-1 and other data center campuses to rapidly boom in the region. 

Why does an outage there ripple across the globe?

Today, AWS data centers exist all over the world, but many tech companies heavily rely on AWS’ US-East-1 region because of its popularity and wide range of services. 

In other words, when Amazon gets a new customer and directs its data processing toward one of its many, many campuses — US-East-1 is usually the default option. Major customers include airline giants, news websites, online shopping platforms, streaming services and nearly every other major corner of the internet. 

On top of those individual customers, US-East-1 runs global infrastructure that power even more parts of the web, too. The campus powers internet traffic controller Route 53 Public DNS NS secure connection service Amazon CloudFront, two backbones of web browsing.

So, even if an online service purposely avoids processing its data via US-East-1, it’s still a central hub for data processing that may be contracted out somewhere else in the supply chain. And one disruption has ripple effects. 

If you buy apples at the market, but there are no trucks to get them from the farm to the market, that market no longer has apples to sell you. If you want to open Netflix to watch TV today, but US-East-1 has an outage, there’s no way to route the show to your screen. 

What services and platforms broke?

A non-exhaustive list of impacted services includes: 

  • Amazon and its subsidiaries
  • AT&T
  • Canva
  • Coinbase
  • Delta Airlines
  • Disney+
  • Facebook
  • Flickr
  • Fornite
  • Hulu
  • McDonald’s
  • Perplexity
  • Playstation
  • Roblox
  • Roku 
  • Signal
  • Snapchat
  • United Airlines
  • Venmo
  • YouTube

What did AWS say caused this?

The company blamed an “underlying DNS issue” for the US-East-1 outage. 

A hiccup in Amazon’s DynamoDB, which manages database at US-East-1, disrupted the links to the servers that power internet connectivity. In other words, the tech that usually translates domain names into IP addresses that computers can process cut out for a bit. 

The outage does not appear to be linked to a cyberattack. If you’re still experiencing lags, AWS recommends flushing your DNS cache.

How often do outages like this happen? Is it likely to happen again?

Data center outages like today’s US-East-1 incident are relatively common, and very likely to happen again. 

In 2017, a US-East-1 outage took down what felt like most of the internet, the 2021 incident is still cited as the biggest AWS distruption in history and a 2023 error took several services offline. It’s not just AWS, either. All data processors are subject to outages whether from power disruptions, cyberattacks or backend errors.

Businesses can shell out cash to avoid this as much as possible, often by running redundant processors to have a backup available if the main one goes out or hosting across multiple different regions under the same provider. But that’s expensive, and some disruptions are always to be expected. 

The internet relies on data centers to function and as increased AI development demands high amounts of compute power, we can expect more of them to hit the grid.