• by Laurie Sullivan , Staff Writer @lauriesullivan, March 3, 2017

Amazon points to human error for the disruption in its cloud services resulting in widespread outages and glitches for Internet-connected devices from companies like Nest, as well as advertisers, news sites, and government services. But with millions -- if not a billion -- of business dollars dependent on AWS, doesn't it make sense to have automated checks and balances?

The blog post -- published Thursday by Amazon about Simple Store Service, also known as Amazon S3 -- describes the human error as an incorrect command that removed a larger set of servers than intended.

The disruption in Amazon S3 lasted for more than 3.5 hours and impacted sending and receiving data for clients such as Apple, Adobe, EB Games, Intuit, Periscope, Live Nation, and Snap.

Trip Chowdhry, managing director at Global Equities Research, believes that 3.5 hours translated into five hours for Amazon's clients because they had to restart their own services once Amazon fixed the glitch.

Amazon's outage will have about a 2% immediate negative impact on the company's first-quarter revenue results, Chowdhry wrote in a research note. And while it's very difficult to pinpoint each loss by Amazon's clients from the disruption, Chowdhry told SearchBlog that ad-serving companies like AdRoll and service companies like Apple or Adobe that rely on cloud services could also see between a 1% and 2% negative impact on their own earnings for the month of February.

AdRoll said it uses AWS to serve 50 billion daily ad impressions. 

Chowdhry said companies will separate their revenue from services to determine the impact.

He also wrote in the research note that "Apple, Adobe, Netflix have negotiated individual SLA penalties with AMZN-AWS, which range from 10% to 30% for Service Credits, if SLA's are not met.  We don't know the exact amount."

Amazon explains in a blog post how the company will make several changes to improve the recovery time of its key S3 subsystems.

"During this event, the recovery time of the index subsystem still took longer than we expected," Amazon wrote. "The S3 team had planned further partitioning of the index subsystem later this year. We are reprioritizing that work to begin immediately."

Amazon reported during its fourth-quarter earnings report that enterprise customers have committed to migrate "tens of thousands of applications" to AWS. Those companies include Workday, Capital One, Matson, McDonald's, Financial Industry Regulatory Authority, and Enel.