What Happens When AWS Goes Down: Regions, Zones and the Dashboard
When Amazon Web Services has a problem, apps that look unrelated can fail at the same time because they rent computing, storage and databases from the same provider. AWS splits its infrastructure into Regions, and a fault in one Region, such as N. Virginia (us-east-1), affects customers who depend on that Region. Apps built to run in several Regions can stay up.
The status of AWS itself is on the AWS Health Dashboard. Our live Amazon Web Services status page tracks it next to user reports. This guide covers how AWS is organized, why failures spread, and how to read official reports.
What are AWS Regions and Availability Zones?
A Region is a separate geographic area, and each Region is made of Availability Zones. AWS documentation describes an Availability Zone as one or more discrete data centers with independent power, networking and connectivity within a Region. AWS states that each Region has at least three Availability Zones and that zones are spaced apart to avoid correlated failures while staying close enough for fast replication.
A company can spread an app across zones inside one Region and still go down if the Region has a shared problem. Zone names like us-east-1a can map to different physical sites in different accounts, which is why AWS offers zone IDs.
Why do so many apps fail at the same time?
Apps share building blocks. A website might use one AWS service for databases, another to start servers, another to sign users in, and a content network in front. If a core service in a Region fails, everything that relies on it can fail with it, even if the app’s own code is fine.
AWS’s own account of the October 2025 event shows how failures chained. In its post-event summary, AWS says the root cause was a latent race condition in the DynamoDB DNS management system that left the DNS record for the regional endpoint empty, so applications could not reach DynamoDB there. AWS then lists later impact on other services in the same Region, including EC2, Network Load Balancer, Lambda and container services. DNS is the system that turns names into addresses; see DNS errors explained.
What did AWS say about the October 2025 outage?
The report is titled “Summary of the Amazon DynamoDB Service Disruption in the Northern Virginia (US-EAST-1) Region,” published at aws.amazon.com/message/101925. It covers October 19 and 20, 2025. AWS gives the overall event window as 11:48 PM PDT on October 19 to 2:20 PM PDT on October 20, and lists separate windows by service. We cite only AWS’s own figures and do not add estimates of how many users were affected, since the report does not.
How do I read the AWS Health Dashboard?
The AWS Health Dashboard has a public Service health view that needs no account, at health.aws.amazon.com/health/status. AWS documentation says it shows public events not specific to an account. Signing in shows account-specific events and an event log of the past 90 days.
- Open the public Service health page.
- Look for the Region your app is in, such as N. Virginia.
- Find the service by name and open any event for its updates.
- Remember the page covers AWS’s services, not the apps that use them. A consumer app can fail while its dashboard row looks normal if the issue is the app’s own.
Update wording follows the same pattern as other status pages; see how to read a status page.
What can I do as a user during an AWS outage?
Very little can be done from your side, and the main job is to avoid making things worse. Do not repeat payments, reinstall apps or give codes to anyone offering “support.” Check the app’s own status page, wait, and check back later. The checklist is in what to do during an outage. For a quick test of whether the problem is local, see is it down or just me.
| Term | Meaning |
|---|---|
| Region | Geographic area with several Availability Zones |
| Availability Zone | One or more data centers within a Region with independent power and networking |
| us-east-1 | The N. Virginia Region |
| Service health | Public dashboard of AWS events |
| Post-event summary | AWS’s written account of a major event |
How can companies avoid depending on one Region?
Companies can run their service in more than one AWS Region, keep copies of data in other Regions and test their recovery plans. That costs more and is harder to build, so many do not. Readers cannot control this, but it explains why some apps recover quickly while others wait for AWS to fix its side.
Every service in AWS’s report had its own impact window, which is why recovery looked uneven to users. In the October 2025 summary, AWS lists DynamoDB, EC2, Network Load Balancer, Lambda, ECS, EKS, Fargate, Amazon Connect, the AWS Management Console sign-in, Redshift and AWS Support Center with different start and end times. Some apps came back as soon as their database did; others waited for server launches or load balancing to recover.
Bottom line
An AWS outage means a provider that many companies share has a fault, which is why unrelated apps fail together. The facts of any event come from AWS’s Health Dashboard and its post-event summaries.
FAQ
Does an AWS outage mean the whole internet is down?
No. It affects services that depend on the affected AWS Region or service. Other sites keep working.
Why is us-east-1 mentioned so often?
It is the N. Virginia Region, one of AWS’s Regions. AWS’s October 2025 summary concerned that Region.
How long do AWS outages last?
It varies by event, and AWS publishes timings in its post-event summaries. We do not generalize.