One of Amazon Web Services’ most popular features was on the fritz Tuesday.
The cloud-computing business’ S3 service, which provides on-demand storage for companies large and small, began experiencing “increased error rates” at its North Virginia site this morning around 11:45 a.m. PT.
The issue was resolved by 2:08 p.m. PT, with Amazon saying its S3 service was “operating normally.”
Websites relying on certain S3 servers weren’t able to access some or all the information they stored with AWS, resulting in a handful of sites experiencing slow loading times or an inability to load at all. Some of those troubled sites included Quora, Imgur, IFTTT, Giphy and Slack.
Ironically, the sites Is it down right now? and Down Detector, used to check on internet outages, were also having hiccups.
“This is significant because S3 is such a core service to AWS,” Dave Bartoletti, Forrester’s public cloud expert, said. “Virtually everyone who uses AWS uses S3.”
S3, which stands for Simple Storage Service, was first introduced in 2006 and is one of AWS’ oldest services. It allows companies to store all kinds of data, such as images for their websites or database backup information. The fact that it faced such noticeable problems was surprising, Bartoletti added, because it’s typically “incredibly solid.” According to AWS, S3 is designed to deliver “99.999999999% durability.”
It’s unclear what caused the S3 problem, but AWS regularly offered updates on its efforts to bring back service as the problem dragged on for a few hours.
“We are working hard at repairing S3, believe we understand root cause,” AWS said in a message at 11:35 a.m. PT, “and are working on implementing what we believe will remediate the issue.”
An AWS spokesperson provided the same statement and didn’t have any additional details.The AWS Twitter page and the AWS Service Health Dashboard offered some additional information.
Forrester’s Bartoletti predicted the problem involved a software failure, adding that there has been no indication so far that the issue related to a physical server problem, such as a fire, or a hack. He hadn’t spoken directly with AWS about the S3 problem.
The S3 issue illustrates how a more connected internet can now have more complicated and noticeable problems when things go wrong. Many companies have turned to cloud services like AWS and rival Microsoft’s Azure to avoid having to buy and update their own servers. These cloud offerings have become lucrative, especially for Amazon, which is the top cloud services provider. However, when things don’t work correctly for these cloud providers, it can have a big impact on the internet.
Throughout the day, Giphy had a few hiccups, plus IFTTT and Nest services. Imgur and Adobe services also reported downtime due to the S3 problem. A handful of publications reported (via their Twitter accounts, of course) issues creating content.
With so many busted things around the web, it was shaping up to be a long day. Of course, some sites had fun with it:
Netflix remained working though. (Not that you’re looking for a way to avoid work.)
First published Feb. 28, 11:28 a.m. PT.
Update, 2:30 p.m. PT: Adds Amazon’s report that S3 service recovered.
Update, 1:24 p.m. PT: Adds more details throughout.