Amazon has disclosed that a bug in its automation software caused the widespread AWS outage this week, which disrupted thousands of websites and applications, from banking services to smart beds. The company published a detailed post on Thursday outlining the cascading events that led to the hours-long disruption.
The outage affected DynamoDB, AWS's database system, due to a latent defect in its automated DNS management system. DynamoDB maintains hundreds of thousands of DNS records, and the automation software failed to repair an empty DNS record for the US-East-1 region in Virginia, requiring manual intervention.
AWS has disabled the DynamoDB DNS planner and enactor automation worldwide while it fixes the underlying conditions and adds extra protections. The issue also caused outages for other AWS tools, impacting platforms such as Signal, Snapchat, Roblox, Duolingo, banking sites, and Ring doorbell.
According to Downdetector, over 8.1 million problem reports were logged globally, affecting around 2,000 companies. Services were restored within hours, but the outage highlighted vulnerabilities in internet infrastructure.
Dr Suelette Dreyfus, a computing lecturer at the University of Melbourne, noted that the incident underscores dependence on a few cloud providers, eroding the internet's original resilience. Eight Sleep, a smart bed company, apologised and rolled out a Bluetooth update to allow offline control of critical functions.
