Amazon Web Services Outage History

The outage began on 1:20 AM UTC in the AP-NORTHEAST-3 region and affected launching new EC2 instances and creating or modifying EBS volumes within a single Availability Zone (apne3-az3). Connectivity to existing instances remained unaffected, and the impact was limited to Osaka services in this region. Several AWS components in Osaka were listed with degraded performance during the incident, indicating a partial outage rather than a regional-wide disruption. The event spanned multiple services and was localized to a single AZ.

Over the course of the incident, AWS issued seven status updates as engineers diagnosed the issue and implemented mitigations. The root cause was identified in an underlying EBS volume subsystem related to EC2 launches and EBS operations, and mitigations were pursued across parallel paths. AWS recommended customers launch new EC2 instances in other Availability Zones while recovery efforts continued. By 3:35 AM UTC, affected Osaka components began recovering and returned toward normal operation across the listed services.

An outage in AWS Identity and Access Management produced increased API error rates. Customers reported issues creating or viewing IAM roles, resulting in degraded IAM functionality across affected accounts. The incident began at 12:50 PM UTC and was resolved by 1:16 PM UTC, lasting about 26 minutes.

The AWS status page showed IAM moving from a degraded state (warn) back to healthy (up) after the incident. The provider issued 2 status updates during the outage.

The AWS EU-NORTH-1 region (Stockholm) experienced a networking issue that caused increased error rates and latencies across multiple AWS services. The problem affected intra-region traffic and most of the impact centered on eun1-az3, with connectivity to and from the region largely intact but many services showing slower responses or timeouts. The outage began on Feb 13, 2025 11:40 PM UTC and extended for several hours as engineers worked to mitigate the issue.

During the incident, AWS published numerous status updates detailing mitigation efforts and affected components. Early signs of recovery appeared around Feb 14, 2025 12:46 AM UTC and full recovery was achieved by Feb 14, 2025 4:05 AM UTC, though some requests may still timeout or be throttled during the transition.

An outage briefly affected Amazon Web Services' Bedrock API usage of Claude 3.5 Sonnet v2 in the US-WEST-2 region. The impact was limited to the Oregon / Amazon Bedrock and Oregon / Amazon Q Business components, causing degraded performance for requests hitting the Bedrock API in that region. The issue began at 11:11 PM UTC and was resolved by 11:45 PM UTC, spanning about 34 minutes. AWS issued four status updates during the incident, noting an investigation into the increased error rates and a subsequent restoration of service. The geographic scope remained regional to US-WEST-2.

Amazon CloudFront experienced degraded performance with intermittent 404 NoSuchBucket errors for distributions using Amazon S3 as the origin. The issue began at 10:20 AM UTC and was resolved by 11:45 AM UTC, lasting about 1 hour and 25 minutes. Users accessing assets via CloudFront may have encountered intermittent errors when the origin was S3-backed.

Amazon CloudFront issued two status updates during the incident. The first update acknowledged increased error rates and that the team was actively investigating, while the second update confirmed the issue had been resolved and normal service had resumed. Overall, only a small number of user reports were observed during the event, indicating limited data points outside of CloudFront's own monitoring.

AWS Management Console in the AP-SOUTHEAST-5 region experienced degraded performance with increased error rates, and AWS WAF showed degraded status in the same region. Users attempting console actions or WAF configurations could have encountered errors or delays, potentially impacting deployments and security rule updates. The incident began with the AP-SOUTHEAST-5 region entering a degraded state at Jan 16, 2025 12:50 AM UTC and was resolved by Jan 16, 2025 3:16 AM UTC.

The AWS Identity and Access Management (IAM) control plane experienced increased error rates affecting IAM Control Plane API requests, causing intermittent failures for IAM operations while authentication and authorization for existing users, credentials, roles, and policies remained unaffected.

The issue began at 9:46 PM UTC and was resolved by 10:35 PM UTC, lasting about 49 minutes. AWS posted two status updates during the incident describing the degraded IAM control plane and the subsequent restoration.