Amazon Web Services outage history

An EU-NORTH-1 EC2 API outage caused elevated error rates and failed instance launches, with some users seeing API errors and 503 responses when launching new EC2 resources. AWS noted that existing instances were generally not affected and that management console actions could also return server errors. The issue began around 9:06 PM UTC and early updates indicated that EC2 APIs were experiencing increased errors in the region, potentially impacting services that rely on EC2 for provisioning or API access.

By 9:25 PM UTC, AWS confirmed the broader impact with increased EC2 API errors and indicated that other AWS services that perform instance launches or call EC2 APIs could be affected. The situation included degraded performance and latency across related services, with multiple EU-NORTH-1 components entering a degraded state. In addition, around 9:55 PM UTC, AWS highlighted damage in the Middle East (ME-CENTRAL-1) region due to regional conflict and urged customers to migrate workloads to other Regions and restore from backups where possible.

Recovery signs began around 9:30 PM UTC as services started to show early improvement, and by 1:45 PM UTC many affected components in Bahrain/UAE and other regions had recovered or returned to up. User reports came from multiple regions, including Europe, the Middle East, and parts of Asia and North America, reflecting a regional pattern rather than a globally uniform outage. Overall, the incident produced widespread degraded performance across EC2-dependent workflows for most of the event window, with a gradual restoration as services progressed toward normal operation.

The AWS Direct Connect service in the US East (N. Virginia) region experienced a brief connectivity loss that affected customers using Direct Connect to reach AWS services. The outage began on 12:40 AM UTC and was resolved by 2:15 AM UTC, lasting about 1 hour and 35 minutes. The issue affected the N. Virginia / AWS Direct Connect component, causing degraded network paths for some customers. The AWS status page issued two updates during the incident, initially noting investigation into connectivity loss and later confirming restoration. The impact was localized to Direct Connect users accessing US-East-1, with limited user reports during the event.

The outage began on 9:20 PM UTC when Amazon CloudFront DNS resolution failures started causing increased error rates and latencies for several CloudFront distributions. Route 53 delays also slowed propagation of changes, amplifying disruption for users accessing content via CloudFront. Users reported outages and errors across multiple services, including 403 responses for certain actions. The incident lasted roughly 7 hours and was resolved by 4:20 AM UTC as CloudFront returned to normal operation.

An incident in the AWS EU-West-1 region (Ireland) led to degraded performance with timeouts when pulling container images. User reports began around 4:30 PM UTC, and AWS acknowledged elevated latencies for network change propagation in EU-WEST-1 at Jan 28, 2026 5:16 PM UTC. The issue impacted multiple Ireland-based components, including EC2, AWS Fargate, Amazon ECS, and AWS WAF.

AWS issued 3 status updates during the incident, at Jan 28, 2026 5:16 PM UTC, Jan 28, 2026 5:25 PM UTC, and Jan 28, 2026 9:35 PM UTC. All affected components returned to normal by Jan 28, 2026 9:35 PM UTC, marking an outage duration of about 5 hours and 5 minutes with region-specific impact to Ireland (eu-west-1).

From Oct 28, 2025 10:40 PM UTC, the US East-1 (N. Virginia) region experienced degraded performance with increased error rates and latencies for EC2 instance launches in the use1-az2 availability zone. Some requests were throttled in this AZ, affecting EC2 launches and related operations across multiple AWS services. The impact extended to many services in the region, including EC2, Fargate, ECS, EKS, DataSync, CodeBuild, EMR Serverless, App Runner, and AWS Batch, among others. User reports indicate broad geographic activity, suggesting widespread regional impact. The outage lasted about 7 hours and 20 minutes and began to improve as mitigations were applied; recovery was largely underway by Oct 29, 2025 6:00 AM UTC.

AWS communications tracked the incident through multiple status updates noting increased error rates and latencies and ongoing throttling in use1-az2 as mitigations were deployed. As stability returned, services progressively recovered and throttling was lifted for many components. By around 06:00 UTC on 2025-10-29, most affected services had recovered and throttling in the region was subsiding. Existing instances were not affected, and AWS indicated throttles would be removed as confidence in stability grew.

The outage began at 11:10 AM UTC in the US-EAST-1 region, with AWS reporting increased error rates and latencies across a broad set of services. Customers experienced issues launching EC2 instances and processing messages with Lambda and SQS, and workloads relying on compute, data services, and analytics were affected. The impact appeared region-wide, impacting dozens of services including EC2, SQS, Lambda, OpenSearch Service, CloudWatch, DynamoDB, and more. The outage continued for many hours as AWS worked to restore functionality.

The outage began around Oct 20, 2025 7:40 AM UTC when AWS reported increased error rates and latencies across the US-EAST-1 Region. Core services including DynamoDB and EC2 showed degraded performance, impacting login, API requests, and console access for users in and around the region. User reports surged in the early window, with noticeable outages and degraded functionality reported between 7:30 AM UTC and 7:45 AM UTC UTC as individuals and organizations tried to access AWS services from multiple geographies.

The AWS Identity and Access Management (IAM) APIs experienced increased error rates and latencies, causing intermittent delays for authentication and other IAM operations. The incident began at Sep 27, 2025 3:30 AM UTC and was resolved by 4:00 AM UTC, lasting roughly 30 minutes. AWS reported investigation into the elevated error rates and latencies on the IAM APIs, with the IAM component transitioning from a degraded state to normal as a result of mitigation efforts.

The IAM service status page showed the issue as a degraded condition before returning to normal, reflecting the temporary impact across IAM-related functionality.

An incident in the EU-NORTH-1 region caused elevated API error rates and latency for Amazon SQS, with the SQS ReceiveMessage API for queues using Server-Side Encryption being particularly affected. The issue also impacted some APIs and workflows across other AWS services in the region, leading to degraded performance for applications relying on SQS and related services. The incident began at Sep 25, 2025 8:25 PM UTC and persisted for several hours as engineers investigated and applied mitigations. The provider issued 15 status updates during the incident, reflecting ongoing analysis and gradual mitigation. The outage progressed toward recovery through the night, with some services returning to normal by Sep 26, 2025 5:30 AM UTC and remaining services staged for full restoration.

AWS Glue in the US-East-1 (N. Virginia) region and Amazon EventBridge experienced degraded performance with increased error rates. The issue began on 9:25 PM UTC and affected Glue job runs and event-driven workflows. The problem lasted until 10:05 PM UTC when services returned to normal.

The AWS status page published two updates during the incident, detailing the investigation and restoration progress. The impact was regional to US-East-1 and affected two components: N. Virginia / AWS Glue and N. Virginia / Amazon EventBridge. The outage duration was roughly 40 minutes.