A power outage at an Amazon data centre in Northern Virginia caused a widespread outage for Amazon Web Services (AWS).
The outage affected several customers, including Slack, Epic Games, EA, Imgur and Trello.
Users began reporting problems with AWS on Downdetector at around 8 a.m ET. Other services also saw a spike in error reports at the same time.
On the AWS status page, Amazon noted that the power outage increased launch failures and networking connectivity issues in a single Availability Zone in the US-EAST-1 Region. Other Availability Zones within the US-EAST-1 Region were not affected by this issue.
Although Amazon restored most services in about two hours, some issues still remain as of 1:30 p.m. ET. AWS recommends that customers relaunch affected EC2 instances within the Availability Zone to restore services.
But AWS’s problems aren’t over. Around 12:30 p.m. ET, new issues with AWS Single Signon have emerged, also at the North Virginia site.
This outage marks the third one for AWS this month. The biggest one occurred on Dec. 7 when an error in the AWS API crippled Amazon’s services, Disney+, Tinder and various other services hosted on AWS. That outage lasted almost 10 hours.
Update:
[3:10 p.m.] AWS is now experiencing increased error rates for customers using Directory Services AD connector or Managed AD with Amazon SSO in the US-East-1 region. Amazon has identified the cause to be related to Directory Services. Resolution efforts are ongoing.