On Monday, October 20, 2025, Amazon Web Services experienced a major incident centered in its US-EAST-1 region. The problem began in the early morning hours Eastern Time and manifested as DNS resolution trouble plus a fault tied to the internal health monitoring of network load balancers—arcane details, but the business consequence was simple: many apps couldn’t reliably log people in, move data between services, or complete purchases. By evening, AWS said services were back to normal, though some queues took longer to drain.  

The blast radius was wide because so much of the modern internet runs through a handful of cloud regions. Snap, Reddit, Zoom, Venmo, Fortnite, and Signal were among the brands reporting disruption; Amazon’s own consumer experience (Alexa, Ring, Prime Video) saw knocks; airlines and banks reported issues in pockets; and the effect was felt from North America to Europe and Asia. Several outlets described this as the most visible cloud wobble since last year’s CrowdStrike fault—not because it lasted forever, but because of how many everyday transactions suddenly felt brittle.  

WIRED’s framing is …