AWS Outage History: The Biggest AWS Downtime Events from 2021 to 2025

Amazon Web Services (AWS) powers some of the largest platforms and applications in the world. Yet even the most dominant cloud provider is not immune to downtime.

Outages at AWS can ripple across the internet, affecting businesses, SaaS platforms, and millions of end users.

To help teams understand and prepare for these risks, StatusGator has been monitoring AWS status updates since 2015. In 2024, StatusGator presented its Early Warning Signals feature. That’s how we were able to alert users of the ongoing service disruptions, even when there’s no official notification. That’s how StatusGator users knew that AWS experienced regional degraded performance on February 12, 2025.

While AWS provides limited transparency during incidents, our platform maintains a 10-year archive of service uptime across thousands of providers, including AWS, Azure, and Cloudflare.

This article highlights some of the most notable AWS outages over the past five years.

October 2025 – Multi-Service Outage in N. Virginia (us-east-1)

On October 20, 2025, Amazon Web Services experienced a major outage in the N. Virginia (us-east-1) region lasting approximately 15 hours. The incident began at 12:11 AM PDT, affecting DynamoDB, EC2, Lambda, CloudWatch, and numerous dependent services across AWS’s ecosystem.

The root cause was traced to DNS resolution issues for DynamoDB endpoints, which triggered cascading failures across EC2 instance launches, Network Load Balancer health checks, and other interdependent components. As a result, hundreds of AWS services were impaired, leading to widespread disruptions for customers worldwide.

Recovery began around 2:24 AM PDT, following the resolution of the DynamoDB DNS issue. However, full restoration of all affected services, including AWS Config, Redshift, and Connect, was not achieved until 3:01 PM PDT.

StatusGator’s Early Warning Signals detected the outage approximately 10 minutes before AWS officially acknowledged it, alerting users to degraded performance across thousands of SaaS services dependent on AWS. During the incident, StatusGator issued over 100,000 notifications, demonstrating how aggregated monitoring enables faster detection and response than relying solely on provider communications.

August 2025 – DynamoDB Streams Outages in GovCloud

On August 13, 2025, AWS experienced simultaneous outages in us-gov-west-1 and us-gov-east-1, disrupting DynamoDB Streams for nearly two hours. The root cause was traced to a planned network configuration change that unexpectedly degraded performance.

• Gov-West (us-gov-west-1): Between 9:36 AM and 11:23 AM PDT, reads from DynamoDB Streams failed, affecting 16 AWS services. Recovery began around 11:44 AM, with full resolution by 12:16 PM.
• Gov-East (us-gov-east-1): A similar issue unfolded from 9:24 AM to 11:31 AM PDT, impacting 14 AWS services. Full recovery was confirmed by 12:15 PM.

Although short-lived, these outages underscore how routine configuration changes can cascade into widespread disruption, particularly for services that act as dependencies across AWS’s ecosystem.

February 2025 – Networking Disruption in Stockholm

A networking fault inside the eu-north-1 region caused widespread errors across EC2, S3, DynamoDB, RDS, Lambda, and more. While traffic into and out of the region was stable, intra-region communication failed for several hours before full recovery.

Amazon Web Services never publicly acknowledged this issue. StatusGator alerted users of the AWS disruption on February 12, 2025, at 2:54 AM EEST. The incident lasted for nearly 8 hours.

December 2024 – Cognito Authentication Failures in us-east-1

On December 13, 2024, AWS Cognito users in us-east-1 faced login failures and “Too many requests” errors due to elevated authentication error rates. The outage lasted 1 hour and 27 minutes, beginning at 4:24 AM. Customers were advised to retry failed requests as the service gradually recovered.

Interestingly, StatusGator detected the issue 32 minutes before AWS officially acknowledged it, once again showing how independent monitoring can provide earlier awareness than relying on provider updates alone.

July 2024 – Kinesis Failure in us-east-1

A flaw introduced during an upgrade to Amazon Kinesis triggered a seven-hour outage in North Virginia. The disruption cascaded into dependent services such as CloudWatch Logs, Lambda, Redshift, and ECS, illustrating the risks of updating AWS’s most critical services in its busiest region.

June 2023 – Regional Outage in us-east-1

Multiple AWS services failed in North Virginia, impacting organizations like The Boston Globe and the Associated Press. Amazon Connect was hit hardest, while Lambda and EventBridge also saw widespread delays.
The disruption caused widespread communication breakdowns for customer support centers relying on Amazon Connect. It reinforced us-east-1’s reputation as both AWS’s most critical region and its most outage-prone.

December 2022 – Connectivity Issues in us-east-2

A 40-minute outage disrupted VPNs and internet connectivity in Ohio. Although brief, the incident highlighted AWS’s limited transparency. No detailed postmortem was provided, and it reinforced the value of external monitoring tools.
Businesses running VPN-dependent applications faced sudden disruptions to remote access and data transfer. The lack of detailed reporting left many enterprises frustrated and seeking third-party visibility.

July 2022 – Power Failure in us-east-2

A short power outage in Ohio caused three hours of instability across EC2 and impacted services like Webex, Okta, Splunk, and BambooHR. Even a 20-minute infrastructure failure had prolonged downstream effects.
This event highlighted how quickly physical infrastructure issues can ripple into cloud service instability. For IT teams, it was a reminder of the importance of multi-region failover strategies.

December 2021 – Multi-Day Disruptions in us-east-1

Two outages within the same week in December caused major issues across EC2, DynamoDB, API Gateway, and more. The first incident lasted over seven hours, followed days later by another wave of disruption, demonstrating how fragile core AWS services can be.
The outages disrupted major consumer applications, including streaming platforms and productivity tools. They also sparked renewed discussions on the risks of centralizing workloads in a single AWS region.

September 2021 – EBS Failure in us-east-1

An eight-hour Elastic Block Store (EBS) issue disrupted EC2 instances and dependent services like RDS and Redshift. What began in a single Availability Zone cascaded into a regional disruption.
Many customers reported data access problems, delayed database queries, and failed backups during the event. The outage highlighted the deep interdependence between AWS storage and compute services.

Why AWS Outage History Matters

For enterprises relying on AWS, these outages are reminders that cloud computing, while resilient, is not infallible. Regional dependencies, service interconnections, and opaque provider communication can all magnify the impact of an incident.

That’s why many IT teams rely on StatusGator to aggregate status pages in real time.

Staying ahead of AWS disruptions is challenging. Especially when official status pages often lag behind real incidents. That’s why many IT and DevOps teams rely on StatusGator to monitor AWS status alongside 6,000+ other cloud and SaaS providers.

With real-time alerts, historical outage data, and early detection of disruptions (often before AWS acknowledges them), StatusGator helps teams respond faster, reduce downtime, and keep critical systems running smoothly.