What stops working when AWS goes down
When Amazon Web Services experiences an outage, the applications that stop working depend entirely on which AWS region failed and which services that app relies on. A Netflix streaming interruption looks different from a Slack outage, which looks different from a Shopify store going offline — because each one uses different AWS infrastructure in different ways. The outage that affects you is the one where your app's backend, database, or content delivery network sits in the region that lost power or connectivity.
AWS operates in geographic regions — US East (N. Virginia), US West (Oregon), Europe (Ireland), Asia Pacific (Tokyo), and others. When one region fails, only apps hosted primarily in that region go down. An app might use multiple regions for redundancy, so it keeps running. An app that uses only one region and has no backup stops immediately.
The services most likely to cause widespread outages when they fail are EC2 (virtual servers), RDS (managed databases), S3 (file storage), and CloudFront (content delivery). If an app's database goes down, the app cannot read or write data. If the servers themselves go down, the app has nowhere to run. If the CDN fails, video and images load slowly or not at all.
Key Takeaways
- Apps stop working only if they run in the AWS region that failed — apps in other regions continue normally.
- Services like Netflix, Slack, Airbnb, and Twitch have all experienced outages tied to AWS failures because they host infrastructure there.
- The impact depends on the specific AWS service that failed: database outages prevent data access, server outages prevent the app from running at all, and storage outages affect file retrieval and media playback.
- Apps with multi-region redundancy stay online during regional outages because traffic automatically routes to a working region.
- You can check AWS's status page in real time to see which services and regions are affected during an outage.
Major apps and services that have been affected by AWS outages
Netflix, Slack, Airbnb, Twitch, Instacart, and Robinhood have all experienced service interruptions tied to AWS outages. In December 2021, an outage in US East (N. Virginia) took down parts of Slack, Chime, and other services for hours. In November 2020, an issue with AWS Kinesis (a data streaming service) affected Twitch, which relies on it for live streaming infrastructure. Shopify stores have gone offline during AWS failures because Shopify uses AWS for hosting and payment processing in certain regions.
The pattern is consistent: the bigger the app, the more likely it uses AWS somewhere in its infrastructure. But size does not may provide impact. A large app with servers spread across multiple AWS regions might lose only a portion of its traffic. A smaller app running entirely in one region might go completely dark.
Not every service outage is an AWS outage. When Slack goes down, it might be Slack's own infrastructure, not AWS. The difference matters because the fix and timeline are different. AWS publishes a status page that shows which services are degraded or down in each region, so you can see whether the problem originates with AWS or with the company running the app.
How to tell if an outage is caused by AWS
Check the AWS Service Health Dashboard at status.aws.amazon.com. This page shows the status of every AWS service in every region in real time. If you see red or yellow indicators next to a service and region, AWS is experiencing an issue there. If the dashboard shows all green but an app you use is down, the problem is with that app's own infrastructure, not AWS.
The dashboard breaks down by region and by service. You might see that EC2 in US East is degraded but RDS in US East is fine, and everything in US West is normal. This tells you that only apps running EC2 servers in US East are affected — apps using RDS or apps in other regions are unaffected.
Many companies also post updates on their own status pages or social media when they experience an outage. Twitter, Slack, and Shopify all maintain status pages where they explain what is happening and when they expect to recover. These pages sometimes specify whether the problem is on their end or with a third-party service like AWS.
Which AWS services cause the biggest outages when they fail
EC2 (Elastic Compute Cloud) is the virtual server service. When EC2 fails in a region, every app running servers there stops. This is the most visible type of outage because it affects the app's core ability to process requests.
RDS (Relational Database Service) hosts managed databases. When RDS fails, apps cannot read or write data, even if their servers are still running. The app stays online but returns errors or blank pages because it has no data to show.
S3 (Simple Storage Service) stores files, images, and videos. When S3 fails, apps cannot retrieve media. Streaming services buffer endlessly, image galleries show broken links, and file downloads fail. S3 is so widely used that its outages often affect multiple unrelated apps simultaneously.
CloudFront is AWS's content delivery network. It caches and serves content from locations near the user. When CloudFront fails, content loads slowly or not at all, even if the origin server is fine. Video platforms and image-heavy sites are hit hardest.
Lambda (serverless computing) runs code without requiring you to manage servers. When Lambda fails, apps that rely on it for background jobs, API responses, or scheduled tasks stop working. This is less visible than EC2 outages but can be just as disruptive.
Why some apps keep running during AWS outages
Apps that stay online during AWS outages use one or more of these strategies: they run in multiple AWS regions, they use multiple cloud providers, or they host some infrastructure outside AWS entirely.
Multi-region redundancy is the most common approach. An app might run in US East, US West, and Europe simultaneously. If US East fails, traffic automatically routes to US West and Europe. The app stays online, though users in the US East region might experience slower speeds because they are now connecting to a farther server.
Some large companies use multiple cloud providers. Netflix, for example, runs on AWS but also uses other infrastructure. If one provider has an outage, the app continues on the other. This is expensive and complex, so only large companies typically do it.
A few apps run entirely on their own servers or on smaller cloud providers. These apps are unaffected by AWS outages because they do not depend on AWS at all. This is rare among consumer apps but more common in enterprise software and specialized services.
How long AWS outages typically last
Most AWS outages last between 30 minutes and 4 hours. The December 2021 US East outage lasted about 7 hours for some services. The November 2020 Kinesis outage lasted roughly 2 hours. Shorter outages (under 15 minutes) often go unnoticed by end users because apps recover before most people realize something was wrong.
The duration depends on what failed and how quickly AWS engineers can fix it. A networking issue might resolve in minutes. A hardware failure might take hours. A cascading failure — where one service's collapse causes others to fail — can take longer to untangle.
During an outage, AWS publishes updates on its status page every 15 to 30 minutes. These updates explain what is happening and give an estimated time to recovery. The estimates are often wrong because the engineers do not know how long the fix will take until they are partway through it.
What you can do during an AWS outage
If an app you rely on is down and you suspect an AWS outage, check the AWS status page first. This takes 30 seconds and tells you whether to wait for AWS to recover or contact the app's support team.
If AWS is the problem, waiting is usually the only option. AWS engineers are already working on it, and there is nothing the app's support team can do to speed recovery. Refreshing the app repeatedly does not help and can actually slow recovery by adding load to the recovering service.
If the app's status page shows no issues but the app is still down, contact the app's support team. They might have information about a problem that is not yet public, or they might be able to route you to a working server or workaround.
For critical work, consider keeping a backup plan. If you rely on a cloud-based service for your business, know which apps have multi-region redundancy and which do not. Some companies keep a secondary tool on standby for outages, or they schedule critical work for times when outages are less likely (though outages can happen anytime).
Frequently Asked Questions
Does an AWS outage affect all apps or just some?
Only apps that run in the affected AWS region and rely on the affected service go down. An outage in US East does not affect apps running in US West or Europe. An app running in multiple regions might lose only a portion of its traffic. The AWS status page shows exactly which services and regions are affected.
Can I see which AWS region an app uses?
Not directly from the app itself. Some companies publish this information in their documentation or status pages. You can sometimes infer it by checking latency or by looking at the app's privacy policy, which might mention data residency. The AWS status page shows which regions are having problems, so if an app goes down when a specific region fails, that app likely runs there.
Why do some apps go down during AWS outages and others do not?
Apps with multi-region redundancy stay online because traffic automatically routes to a working region. Apps running in only one region go down if that region fails. Some apps use multiple cloud providers, so an AWS outage does not affect them. The app's architecture — not its size or popularity — determines whether it survives an outage.
How do I know if an outage is AWS or the app's fault?
Check the AWS Service Health Dashboard at status.aws.amazon.com. If it shows red or yellow indicators, AWS is having problems. If it shows all green, the problem is with the app's own infrastructure. The app's status page (if it has one) can also clarify whether the issue is on their end or with a third-party service.
Should I switch to a different app if it experiences AWS outages?
Outages are rare enough that switching apps is usually not necessary. Most apps experience one or two significant outages per year, often lasting under an hour. If an app goes down multiple times per month, that suggests poor architecture or reliability practices — in that case, switching makes sense. If outages are infrequent, the app is probably fine.