Amazon S3 Recovers from Widespread Disruptions
A wide swath of Amazon Web Services (AWS) applications -- and parts of the Internet in general -- were down or impaired for several hours on Tuesday due to a glitch in the Amazon Simple Storage Service (S3).
The problem began just before 10 a.m. PST on Tuesday and was concentrated in AWS' Northern Virginia (U.S. East-1) region, the company's oldest and the one that is estimated to house the vast majority of its cloud servers.
AWS said on Twitter and on its Service Health Dashboard -- which was also temporarily affected -- that the S3 service running out of the Northern Virginia region was experiencing "high error rates," in turn "impacting applications and services dependent on S3."
AWS posted periodic updates on its Service Health Dashboard throughout the morning and early afternoon. Just after 1 p.m. PST, AWS reported that "S3 object retrieval, listing and deletion are fully recovered now. We are still working to recover normal operations for adding new objects to S3."
After 2 p.m. PST, roughly four hours after the service issues began, AWS reported that the S3 problems had been resolved, though it did not provide an explanation for the cause. "As of 1:49 PM PST, we are fully recovered for operations for adding new objects in S3, which was our last operation showing a high error rate," the company said. "The Amazon S3 service is operating normally."
Affected AWS applications included EC2, Athena, Glacier, Redshift, Cognito, Elastic MapReduce, Inspector, Kinesis, Elastic Beanstalk, WorkMail, WorkDocs, CloudFormation and Lambda, among others.
A long roster of Web sites that rely on AWS applications based in the Northern Virginia region -- including Expedia, Medium, Slack, Imgur, Trello, GitHub, Docker and the U.S. Securities and Exchange Commission -- also experienced problems as a result.
According to analysis by Apica, a Web site testing and monitoring company, more than half of the top 100 online retailers
suffered performance declines of at least 20 percent, with their Web sites taking an average of 43 seconds to load. Some retailers like Target, Nike and Nordstrom saw their load times increase between 500 and 990 percent -- and, in the case of the Disney Store, by over 1,000 percent.
Gladys Rama is the senior site producer for RCPmag.com and senior editor of AWSInsider.net.