- Related: AWS Blames Unplanned Systems Restarts for S3 outage
A wide range of Amazon Web Services (AWS), as well as parts of the Internet, were affected by a glitch in the Amazon Simple Storage Service. The problem started just before 10 a.m. Pacific Standard Time (PST) on Tuesday. It was concentrated in AWS’ Northern Virginia (U.S East-1) region, which is the oldest and where the company estimates the majority of its cloud servers. AWS stated on Twitter and its Service Health Dashboard, which was also temporarily affected, that the S3 service operating out of the Northern Virginia region was experiencing high error rates, which in turn was “impacting applications dependent on S3.” AWS posted periodic updates to its Service Health Dashboard throughout morning and afternoon. AWS reported shortly after 1 p.m. Pacific Time that S3 object retrieval and listing were complete. We are still trying to recover normal operations to add new objects to S3. AWS reported at 2 p.m. PST that the S3 issues had been resolved. However, it didn’t give any explanation. The company stated that they were fully recovered to perform operations for adding new objects to S3 as of 1:49 PM PST. This was the last operation with a high error rate. The Amazon S3 service is now operating normally. Among the affected AWS applications were EC2, Athena and Redshift, Cognito. Elastic MapReduce Inspector, Inspector, Kinesis. Elastic Beanstalk. WorkMail, WorkDocs. CloudFormation, Lambda. A long list of Web sites that rely upon AWS applications based within the Northern Virginia region — such as Expedia, Medium and Slack, Imgur and Trello — also suffered from the problem. Apica, a Web-site testing and monitoring company, found that more than half of the top 100 online retailers experienced performance declines of at most 20%. Their Web sites took an average of 43 seconds for loading. Target, Nike, and Nordstrom all saw load times rise between 500 percent and 990 percent, with the Disney Store seeing their load times rise by more than 1,000 percent.