What Is the AWS Outage?

Why Did It Break the Internet for Millions?

On October 21, 2025, millions of people across the world saw their favorite apps and websites freeze without warning. Snapchat stopped loading, Reddit went blank, and even banks like Lloyds and Halifax faced downtime. The culprit wasn’t a cyberattack or a rogue bug – it was a Domain Name System (DNS) failure inside Amazon Web Services (AWS), the internet’s invisible backbone.

In one of the largest global outages since last year’s CrowdStrike crash, a malfunction at AWS’s US-EAST-1 data centre in Virginia cascaded across the web, crippling more than a thousand platforms – from Fortnite and Duolingo to Zoom and Coinbase. Though services are now restored, the event exposed a hard truth: the modern internet depends on a handful of cloud providers, and when one sneezes, the entire digital world catches a cold.

Image 1: When the backbone of the internet falters, the ripple isn’t virtual – it’s global.

What Is the Story of the AWS Outage?

The Cloud That Connects Everything

AWS isn’t just Amazon’s side business – it’s the digital plumbing of the world. One in every three websites or apps runs on its servers. From storage to databases to streaming, AWS provides the infrastructure that lets the internet breathe.

A Small Glitch, A Big Collapse

The outage began in the US-EAST-1 region – AWS’s oldest and busiest hub. A routine internal issue in the load balancer health monitoring system disrupted DNS resolution for the DynamoDB API, a database service countless apps rely on. When DNS – the ‘phonebook of the internet’ – fails, devices can’t find where to send or receive data. It’s like losing every street sign in a city at once.

From Slowdowns To Shutdowns

By early Monday morning, reports on Downdetector skyrocketed to over 11 million. Apps flickered in and out of service. Amazon throttled parts of its network to isolate the problem, triggering what experts called ‘cascading failures’. It took nearly six hours before AWS declared ‘normal operations’ – but backlogs in services like Config, Redshift, and Connect lingered late into the night.

A Familiar Region, A Familiar Pain

The US-EAST-1 region has seen similar collapses before – in 2021 and 2023. Many developers stick to this region by default, making it a single point of failure for global apps. As the outage spread, businesses from London to Tokyo scrambled for answers.

From Crisis to Questions

While AWS apologised and promised a detailed ‘post-event summary’, experts say this was more than a technical mishap – it was a structural warning. Over-reliance on a few American tech giants (AWS, Microsoft, Google) has made the world’s digital economy dangerously centralised.

Image 2: A small misfire in one corner of the cloud can send shockwaves across the digital sky.

The Great AWS Outage (Oct 20 – 21, 2025) – In Numbers

• ~6 hours: Duration of major disruption + delayed recovery

• 1,000+: Apps and Sites Affected

• 11 million+: User Reports (Downdetector)

• 2021, 2023, 2025: Repeat Region Failures

• >$1 million: Peak Revenue at Risk, per hour for large clients

Historical Note

This isn’t the internet’s first blackout – nor its last. In 2021, Meta’s apps vanished for six hours due to a configuration error. In 2024, a CrowdStrike update crippled hospitals and airports worldwide.

And long before that, Fastly and Akamai glitches briefly silenced news sites and banks.

Each time, the pattern repeats: hyper-connected systems, minimal redundancy, and a domino effect from a tiny mistake.

The AWS incident of 2025 is the latest reminder that digital resilience still lags behind digital growth.

What Is the Story of AWS?

It began quietly inside Amazon’s buzzing corridors in the mid-2000s. Engineers were struggling to manage the company’s own fast-growing data systems – so they built internal tools to store, scale, and serve millions of requests more efficiently.

What started as a practical fix soon revealed a bigger idea: what if every company could rent this invisible infrastructure instead of building its own?

And thus, Amazon Web Services (AWS) was born – a digital utility that turned computing power into a service, much like electricity.

In just two decades, AWS reshaped the global economy. Startups could launch without servers; governments could run portals without local hardware; streaming, shopping, and schooling all began to float in the same virtual sky.

But like any great invention, its strength bred dependency. The more the world moved to ‘the cloud’, the more its fate was tied to AWS’s quiet hum of machines in data centres scattered across the planet.

Today, AWS is the unseen engine behind our clicks and comforts. Yet as the latest outage showed, even a single flicker in this engine can ripple through our lives – from a student unable to submit a project to a bank unable to process a payment.

The story of AWS is the story of modern civilization learning to live in the clouds – and occasionally forgetting that even clouds can fall.

Image 3: From Amazon’s server rooms to the world’s data highways – AWS became the cloud that connects us all.

What Is DNS – and How Can One Failure Paralyse the Web?

Think of the Domain Name System (DNS) as the internet’s phonebook. It translates human-friendly names like bbc.com into machine-readable IP addresses. When DNS fails, your browser doesn’t know where to go – even if the destination (like Snapchat’s servers) still exists.

Monday’s failure wasn’t due to hackers or hardware – it was a simple DNS resolution error that left AWS ‘unable to see where to direct traffic’. In a world where everything from payments to gaming depends on split-second lookups, even a brief DNS blackout can cause digital mayhem.

Image 4: When the web loses its map, even giants get lost.

What Is a Downdetector?

Downdetector is a real-time platform that collects user reports of service failures and visualises them as spikes and heat maps. During this AWS incident, it logged over 11 million reports in a single day – one of the highest in its history – offering a crowd-sourced snapshot of the internet’s collapse.

So, What Is AWS and Why Do So Many Apps Rely on It?

Amazon Web Services is the world’s largest cloud computing platform, powering everything from Netflix streams to Uber rides. Instead of buying their own servers, companies rent Amazon’s – gaining flexibility and global reach. But this convenience comes at a cost: dependence.

When AWS falters, the ripple effect hits finance, healthcare, entertainment, education – practically every sector. It’s the ultimate paradox: the more efficient the cloud becomes, the more fragile it grows.

What Is a DNS Resolution Error?

Every time you open an app, a DNS server ‘resolves’ its name into an IP address. If the DNS can’t complete that lookup – whether from a faulty configuration, hardware failure, or network glitch – the request dies midway. That’s exactly what happened here: AWS’s DNS system lost track of its own API routes.

Why Does US-EAST-1 Keep Making Headlines?

The Virginia data centre – nicknamed US-EAST-1 – is AWS’s largest and oldest facility. It’s also the default region for many global clients. That means when it goes down, the blast radius is enormous.

Experts say AWS’s internal systems – EC2 networks, elastic load balancers, and DynamoDB services – are tightly interlinked. So, an issue in one subsystem can trigger a domino effect across others. The lesson? Redundancy on paper doesn’t equal resilience in practice.

Could This Have Been Prevented?

Yes – at least partially. Engineers have long warned about ‘single-region risk’. Large apps could replicate their systems across multiple AWS regions or even multiple cloud providers (known as multi-cloud resilience). But this redundancy costs money, and many startups cut corners to save it.

As Cornell professor Ken Birman told Reuters: “When people cut costs and skip that last step – protection against an outage – they’re the ones who ought to be scrutinised later.” The cloud isn’t invincible; it’s just someone else’s computer.

What Is Multi-Cloud Resilience?

Multi-cloud resilience means spreading your infrastructure across more than one cloud provider – for example, running some services on AWS, others on Microsoft Azure or Google Cloud. If one fails, the others keep you alive. It’s the digital equivalent of not putting all your eggs in one basket – but few companies do it well because it’s complex and costly.

Who Bears Responsibility – AWS or the App Teams?

Both. AWS runs the infrastructure; clients build on it. Amazon’s internal DNS misstep triggered the outage, but many apps went dark because they had no secondary region or backup. The shared-responsibility model means AWS owns the hardware – developers own their resilience.

Thus, AWS provides the tools for resilience, but it’s up to app developers to use them. Many companies still host their critical systems in a single AWS region to minimise costs and complexity. The result: widespread downtime when that region fails.

So, the shared-responsibility model of the cloud means both sides share the blame – AWS for the fragility of its core systems, and clients for not building strong enough backups.

Will Regulators Step In?

Governments are watching closely.

Experts argue that cloud concentration poses a national-security and economic risk. Around 70% of global cloud infrastructure lies with just three U.S. giants: Amazon, Microsoft, and Google. European officials have already debated building ‘local cloud sovereignty’ to reduce dependency. Monday’s outage may reignite those calls.

When ‘one bridge collapses’, as one analyst put it, ‘the entire digital economy stalls’. The question now isn’t whether AWS can fix its systems – but whether the world can diversify beyond them.

The centralisation poses national-security and economic risks. Europe and the UK have floated the idea of ‘cloud sovereignty’ – building local alternatives to reduce dependency on U.S. tech giants. The AWS outage may turn that debate urgent.

WGF Take – Time to Plant Our Feet Back on the Ground, Not Just in the Clouds?

The AWS outage of 2025 was not just a technical blip; it was a mirror. It reflected how a few corporations now hold the keys to the world’s digital infrastructure. Efficiency, convenience, and global reach have come at the cost of independence and resilience.

We have built a world where virtually everything runs on invisible clouds managed by a few companies. But when those clouds have a hole in them, our daily lives experience thunderstorms – and sometimes, full-scale digital storms.

It’s time to rethink resilience as a public priority, not a budget option. Whether through regional diversity, open-source infrastructure, or cloud accountability laws, the internet’s future depends on grounding its foundations before the next storm.

Because the cloud is meant to lift us – not to keep us up in the air.

Image 5: Every storm in the cloud reminds us – true resilience begins on solid ground.