The AWS Outage: What Happened?

On 20 October 2025, AWS experienced a major service disruption, primarily in its US-East-1 region (Northern Virginia).

AWS’s status page indicated “increased error rates and latencies for multiple AWS Services in the US-EAST-1 Region.”
The fallout was broad: major websites and applications like Fortnite, Snapchat, Duolingo, banking apps (in the UK for example) and Amazon’s own services (shopping site, Alexa, Prime Video) all reported problems.
One news report described it as a “major internet outage” because many services globally were impacted—even though the root was in a single region.
Real-time monitoring sites and outage trackers recorded a spike in user reports for AWS-dependent services.

In essence: a disruption in a cloud‐region that hosts a huge number of services translated into what felt like “the internet is down” for many users.

Why the US-East-1 Region Matters (and Why Its Failure Rings Wide)

Understanding why this one region in AWS can cause such broad pain is crucial. Here are some factors:

The role of US-East-1

US-East-1 (Northern Virginia) is AWS’s oldest and one of the largest and most central regions. Many companies host critical workloads there because of lower latency to the U.S. East coast, large ecosystem, availability of services, etc.
Because of its size and centrality, many downstream services (and even upstream dependencies) are anchored in or route through that region.
As history shows, this region has been subject to major outages before. For example: in 2021, an outage due to network device impairment in US-East-1 disrupted many services.

Cascading effects & dependencies

When a core region has increased error rates or latency, “just” one service failing can cascade. For example, if a shared service (say a database, an identity service, a global table) is downstream of many workloads, its failure impacts many.
Many companies assume “the cloud” gives them great resilience—but if you rely heavily on one region and don’t architect for cross-region redundancy, you are vulnerable.
For end users, if enough big services go offline (banking apps, streaming platforms, social networks), it feels like the internet is down—because many of the things you use disappear.

Network / Internet connectivity entanglement

Some of the disruption stems not from your local ISP, or your home network, but from upstream connectivity and routing issues triggered by AWS’s region. For example, if a global content delivery service, DNS resolver, or API gateway hosted in US-East-1 misbehaves, large parts of the world may see slowdowns or failures.
Moreover, commentators note that the “internet core” has a hub‐and-spoke character: many services route via major hubs, and if a hub fails, large disruptions ensue.

The Root Causes (What We Know So Far)

AWS offered only limited public detail at the time of writing, but we can summarise what is known and what past outages suggest.

Known facts

AWS status said: “We can confirm increased error rates for requests made to the DynamoDB endpoint in the US-EAST-1 Region. This issue also affects other AWS Services in the US-EAST-1 Region.”
The timeline: the event seemed to start early (in U.S. early morning local timezone) and propagated quickly to users globally.
The wide dependency means that when core services degrade, many surface services degrade or become unavailable too.

What past analysis tells us

In previous US-East-1 outages, human error or configuration issues have played a role (for example, in 2017 a human error in a debug command removed more capacity than intended).
Other times, network device failure, power/glass cut, or cooling‐system failure have been triggers.
The mix of compute, storage, network, identity and database services means that failure in one subsystem can ripple across many dependent services.

Why “internet down” feeling

It’s not truly that the whole internet is down—just that many of the services you rely on hit faults.
Because AWS is so heavily embedded in the backend of many platforms, when it has problems, many top‐end user experiences fail simultaneously.
For example: banking app fails to log you in, your streaming service buffer spins forever, your smart home device stops responding → you perceive “something big” is wrong with the web, even if other isolated sites still work.

The Impact: Who Felt It & How

Users globally reported inability to access streaming services, gaming platforms, mobile apps and websites.
Financial services suffered: UK banks reported customers unable to log in, payments delayed.
Core AWS customers and internal Amazon services likewise had issues.
Organisations running on AWS (especially in US-East-1) may have hit increased error rates, timeouts, slowed API responses.

For many end users, this meant: you clicked a link, it hung; you opened an app, it said “error-something”; you tried a bank transfer and it didn’t go through; you couldn’t stream your show.

What This Teaches Us: Key Lessons & Best Practices

For organisations & developers

Don’t treat a “single region” as fully reliable If all your workloads live in one region (especially a heavily used region like US-East-1), a failure in that region can take you offline. Multi‐region architecture is expensive but real if you need high availability.
Understand and map your dependencies It’s not enough to say “we run in region X”. What downstream services (databases, global tables, identity, third-party APIs) do you rely on? If they live in the same region or a single point of failure, you’re exposed.
Plan for degraded mode Design for partial failure. If one region goes down, your failover region should be able to handle at least core functionality—even if at reduced capacity.
Monitor not just availability, but error rates and latencies Often the early sign is “requests are slower” or “error rate increased” rather than a complete blackout. Be alert to these signs.
Test your incident & failover processes When an outage happens you’ll have limited time. Do you have a playbook to shift traffic, spin up in another region, bring down dependencies? Test it.
Communicate to users A big part of damage from incidents is loss of trust. If your service goes down, timely communication helps.

For end-users & ordinary consumers

Recognise that when many apps/websites go down at once it may not be your WiFi or ISP—it could be upstream cloud infrastructure.
Follow status pages of major cloud providers or big apps you use: many publish live updates.
If one site is down, check others. If many major services (especially those you use) are failing, it may be large scale.
For mission-critical personal use (e.g., banking, business), consider keeping alternatives (e.g., offline or alternate service) and not relying on “always online”.

What’s Next: Recovery & What to Watch

AWS engineers are actively engaged in mitigating the issue in US-East-1. You should watch:

The official AWS Health Dashboard for updates.
For your own services, watch your monitoring systems for resumed normal error and latency levels.
For end-users, watch the apps/websites you use—many will publish restoration notices.

Once full recovery is achieved, there will very likely be a post‐mortem from AWS (either publicly or via customer communication) explaining root cause, lessons learned, and preventive actions.

Final Thoughts

While it may have felt like “the internet is down”, what really happened is a major cloud provider region (US-East-1) experienced increased error rates and latency, causing widespread disruption across many dependent services. It highlights how deeply the global web ecosystem depends on a few key infrastructure hubs—and how the failure of one such hub can ripple.

For businesses, it re-emphasises the need for robust architecture and dependency awareness. For users, it’s a reminder that when many services go dark at once, the issue may lie upstream in cloud infrastructure rather than in your local connection.