According to the supplied Wikipedia page, Amazon Web Services suffered a major disruption in the US-EAST-1 region in October 2025 after a time-of-check to time-of-use (TOCTOU) race condition affected DNS management for DynamoDB. The system applied an outdated DNS plan after newer plans had already been cleaned up, deleting endpoint IP addresses and causing widespread service failures. The incident illustrates how automation that validates infrastructure state and later acts on that state can operate on obsolete information when concurrent operations are not safely coordinated.
The failure was architectural rather than a conventional intrusion: a race in AWS’s control-plane workflow allowed stale configuration to overwrite newer state. The source does not identify an executive, disclose outage duration, quantify affected customers or financial losses, or cite an AWS postmortem, regulator, or court finding. It therefore establishes the technical mechanism only through the supplied reference, not an independently documented accountability record. TOCTOU defects have longstanding security implications, including privilege escalation, but no evidence in the source indicates that this AWS incident involved customer-data theft or unauthorized administrative access.
