Cloud giant Amazon Web Services wrestled with a major fault at its US-East-1 data centre region in North Virginia on Wednesday, which caused service outages for major internet properties for around three hours.
The @awscloud Twitter account alerted customers in the morning AEST that the S3 offering was "experiencing high error rates."
Apart from certain Amazon services like Twitch, well-known sites with applications and storage in the US-East-1 region were either completely inaccessible or operating slowly.
Impacted sites included Slack, IFTTT, Quora, Adobe, Signal, Pinterest, Zendesk, Heroku, GitLab, AirBnB, the United States Securities and Exchange Commission, Expedia, and more.
The AWS-hosted downtime detection site isitdownrightnow.com also went offline.
The S3 fault appears to have had a snowball effect: multiple AWS services in the US-East-1 region were disrupted, according to the AWS status page. These include:
- Elastic MapReduce
- Kinesis Firehose
- Simple Email Service
- Simple Storage Service (S3)
- Auto Scaling, and
Several other important cloud services such as AWS Elastic Compute Cloud (EC2), Load Balancing, and File System also showed as having performance issues with increased error rates.
However, the company warned that its dashboard was not functionality properly due to the S3 fault. It advised customers to monitor the banner of the top of the page for updates.
Several hours later AWS said it believed it has identified the root cause of the S3 outage and was working to repair the problem.
It did not provide an estimated time of resolution.
Update 10:30am: AWS said it had resolved the S3 outage at 8.49am AEST.