Over a Thousand Websites Affected! Amazon Cloud Service Suffers Worst Outage in Four Years: Lasting 15 Hours, Potential Losses May Exceed $10 Billion

Around 3:00 PM local time on October 20, Amazon AWS announced that it had resolved a service outage that lasted about 15 hours that day, stating "all AWS services are now operating normally." However, some services (such as AWS Config, Redshift, and Connect) still had message backlogs, which would be processed within the next few hours.

Currently, AWS has not yet provided a detailed report on the cause of the failure. According to data from Synergy Research Group, as one of the world's largest cloud service giants, AWS holds over 30% of the global cloud service market share and operates data centers worldwide.

This outage was triggered by a database network failure, affecting hundreds to thousands of websites and applications globally. On the 20th, around 12:00 AM Pacific Time, the US-EAST-1 availability zone, one of AWS's core nodes, was the first to report "increased error rates and latency." This node, located in Northern Virginia, USA, is the earliest activated and largest primary node, where many global services are deployed by default.

Among the first batch of severely affected websites and applications were Amazon, the chat apps Snapchat and Facebook, as well as the popular game Fortnite and the learning platform Canvas, among others. According to the announcement, AWS confirmed at 12:26 AM on the 20th that the trigger cause was a "DNS resolution issue for regional DynamoDB (AWS's cloud-native database) service endpoints."

According to statistics from the network outage tracking website Downdetector, websites or applications such as the financial services companies Venmo and Robinhood, the cryptocurrency exchange Coinbase, Apple's music and TV products, the AI company Perplexity, the video platform Zoom, the Sony game platform PlayStation, and United Airlines all experienced service interruptions that day. The UK government website Gov.uk and HM Revenue & Customs also encountered problems. Within just two hours after the outage began, complaints from the US region alone exceeded 20,000.