Amazon's AWS experienced a significant outage due to a major DNS failure linked to a race condition within DynamoDB's infrastructure, affecting users globally for over 14 hours. The incident led to the accidental deletion of all IP addresses for the database service's regional endpoint, causing widespread connectivity issues. In response, Amazon has implemented measures to prevent future occurrences and apologized for the disruption caused to customers.

The article discusses a significant 14-hour outage of AWS's us-east-1 region, which affected 140 services including EC2, due to a latent race condition in the DynamoDB DNS management system. The author analyzes the outage's causes and emphasizes the complexity and critical nature of AWS's infrastructure, suggesting that oversimplified explanations do not capture the depth of the incident.