What happens to your apps and services during an AWS outage

When Amazon Web Services (AWS) experiences an outage, any process or website built on AWS infrastructure stops working or slows dramatically. This includes household names like Netflix, Slack, Airbnb, and Spotify — all of which rely on AWS to run their core services. The scope of impact depends on which AWS region failed and whether the company built redundancy into their system.

AWS operates data centers in geographic regions around the world. If the US East region goes down, services in that region fail when ready. A company that only runs in one region loses everything. A company that runs copies of their service in multiple regions may stay online but could lose data or experience slowdowns as traffic reroutes.

The outage also affects services you do not see directly. If a weather app uses AWS to store your location history, or a banking app uses AWS to process transactions, those features stop working even though you are not visiting AWS yourself.

Key Takeaways

  • Any website or app that runs on AWS infrastructure will experience downtime or degraded performance during an AWS outage, regardless of how well-known the service is.
  • Companies that run their service in only one AWS region lose all functionality; companies with multiple regions may stay partially online but often experience slowdowns.
  • You may not know a service uses AWS until it fails — the connection is invisible to the end user.
  • AWS publishes a status page showing which services and regions are affected, updated in real time during an outage.
  • Outages typically last from minutes to several hours; major incidents affecting multiple regions are rare but have happened multiple times per year historically.

How to tell if an outage is AWS or the service itself

When a service stops working, the first question is whether AWS is down or whether the company running the service has their own problem. Check the AWS Service Health Dashboard at status.aws.amazon.com. This page shows the status of every AWS service in every region, updated every few minutes during an incident.

If the dashboard shows green across all services and regions, AWS is not the problem — the outage is on the company's side. If you see red or yellow warnings, AWS is experiencing issues. The dashboard also shows a timeline of what failed and when it was resolved.

You can also check the company's own status page. Most major services maintain a public status page (often at status.companyname.com) that explains what is broken and what they are doing about it. During an AWS outage, they will usually post a message saying they are investigating or that they have identified the AWS issue.

Types of applications most vulnerable to AWS outages

Applications that run entirely on AWS with no backup infrastructure are most vulnerable. This includes many startups, smaller SaaS companies, and newer services that have not yet built redundancy. A single-region deployment — where all servers run in one geographic location — means total failure if that region goes down.

Services that depend on AWS databases like RDS or DynamoDB are also heavily impacted because the database itself becomes unreachable. Even if the company has backup servers in another region, they cannot access their data if the primary database is down.

Applications that use AWS-specific services like Lambda (serverless computing), S3 (file storage), or CloudFront (content delivery) are tied to AWS more tightly than applications using only basic compute and storage. If those services fail, there is no quick workaround.

Large enterprises and established tech companies often have less impact because they run on multiple cloud providers or maintain their own data centers alongside AWS. Netflix, for example, runs on AWS but has built systems to handle regional failures gracefully.

What happens to your data during an outage

Your data does not disappear during an outage. AWS stores data redundantly within each region, so even if one server fails, your information is safe on another server in the same region. The problem is that the process cannot reach the data while the outage is happening.

If a company has set up backups in a different region, they can restore service faster by switching to the backup region. If they have not, they have to wait for the primary region to come back online.

Data loss during an outage is rare unless the company was already in financial trouble or had disabled backups. The bigger risk is data inconsistency — if an process was in the middle of writing information when the outage hit, that write might be incomplete. Most modern applications handle this by rolling back incomplete transactions, but it depends on how the company built their system.

How long AWS outages typically last

Most AWS outages last between 15 minutes and 2 hours. AWS engineers can usually identify and fix the problem quickly because they have detailed monitoring and automated recovery systems. A 15-minute outage might affect millions of people but is usually resolved before most users even notice.

Larger outages that affect multiple services or multiple regions are rarer but have happened. In December 2021, a major outage in the US East region lasted about 7 hours and took down a significant portion of the internet. In 2019, a similar incident lasted 4 hours.

The duration depends on what failed. A networking issue might be fixed in minutes. A database corruption might take hours to diagnose and repair. AWS publishes a post-incident report after major outages explaining what went wrong and what they changed to prevent it.

What you can do if a service you rely on is down

If you are using a service that depends on AWS and it stops working, your options are limited. You cannot fix an AWS outage yourself. The best approach is to check the AWS status page and the service's status page to confirm the outage is real and not a local problem on your end.

If you are a paying customer of the service, contact their support team. They can tell you whether they are aware of the issue and when they expect to be back online. If you are a free user, you usually have no direct support channel.

For critical services — banking, email, work communication — consider whether you need a backup. Some people use multiple email providers or keep a second phone number with a different carrier. For work tools, knowing which services your company depends on helps you plan for outages.

If you are building an process yourself, the lesson is to design for failure. Use multiple AWS regions, back up your data to a different cloud provider, or use a hybrid approach with your own servers plus AWS. The cost of redundancy is lower than the cost of being offline.

Why AWS outages affect so much of the internet

AWS holds roughly 32% of the cloud market share, with Microsoft Azure and Google Cloud as distant competitors. This concentration means that a single AWS failure can affect a huge portion of the internet at once. If AWS US East goes down, services across finance, media, retail, and communication all fail simultaneously.

This is a trade-off. AWS is reliable, well-engineered, and cheaper than building your own data centers. Companies choose AWS because the uptime is high and the cost is low. But the choice also means that when AWS fails, many companies fail together.

Regulators and security experts have raised concerns about this concentration. If AWS were to experience a catastrophic failure, the economic impact would be enormous. AWS has responded by improving redundancy, adding new regions, and publishing transparency reports about their reliability.

Frequently Asked Questions

Can I do anything to protect myself from AWS outages?

If you are a user of a service, you cannot prevent outages, but you can prepare. Keep phone numbers and email addresses of important contacts written down. For work, know which tools your team depends on and have a backup plan. If you are building an process, design it to run on multiple cloud providers or regions so one outage does not take you offline.

Will AWS refund me if their outage costs me money?

AWS offers service credits for outages that exceed their stated uptime may provide, but the credit is usually small — often 10% to 30% of your monthly bill. You have to request the credit; AWS does not issue it automatically. Read your service agreement to understand what you are may have access to to.

How do I know if AWS is having problems right now?

Visit status.aws.amazon.com to see real-time status of all AWS services and regions. You can also subscribe to notifications so AWS emails you when an outage starts and when it is resolved. Most services also have their own status pages you can check.

Do all cloud providers have outages?

Yes. Microsoft Azure, Google Cloud, and smaller providers all experience outages. AWS outages are more visible because more services run on AWS, but no cloud provider is immune to failures. This is why companies building critical systems often use multiple providers.

What is the difference between an AWS outage and a service outage?

An AWS outage means AWS infrastructure itself is down — servers, networking, or storage. A service outage means the company running the service has a problem with their code, configuration, or their own systems. AWS outages affect many services at once; service outages affect only that one company.