AWS us-east-1 outage (December 2021)

The AWS (Amazon Web Services) us-east-1 outage of December 2021 is the partial failure of AWS's oldest region, triggered by an automated scaling activity that congested its internal network for about seven hours. On December 7, 2021, at around 7:30 a.m. Pacific time, an automated activity to scale the capacity of an AWS service hosted on the main network triggered unexpected behavior from a large number of clients inside the internal network. Connections surged, the devices linking the internal network to the main network became congested, and latency spiked. That internal network carries foundational services: internal DNS (Domain Name System), monitoring, part of the control planes. AWS teams lost their own diagnostic tools, which slowed remediation, and the public Service Health Dashboard lagged behind events for much of the day. EC2 (Elastic Compute Cloud) instances already running kept running; what failed were API (application programming interface) calls, launches and services that rely on the control plane. Recovery came in the early afternoon, roughly seven hours later. The services hit, from Disney+ to Ring doorbells and Amazon's own delivery logistics, showed that AWS's oldest region concentrates dependencies its customers cannot see. For the underwriter, the event falls under non-malicious failure of a service provider, with no physical damage, and its duration compares directly with policy waiting periods. Seven hours lands just under the eight to twelve hour waiting periods common in the market, which makes it a test of wording more than of capacity. Accumulation modelers cite it alongside the 2017 S3 (Simple Storage Service) outage and the October 2025 one.