Americas

  • United States
michael_cooney
Senior Editor

Happy Amazon Crash Day

News Analysis
Jul 17, 20182 mins
Cloud ComputingData Center

Amazon web servers weren’t up to the Prime Day deluge

amazonprimeday2018
Credit: Amazon

On one of the biggest shopping days of the year for Amazon.com the company’s web site crapped out intermittently for hours yesterday.

Instead of Prime Day purchases, many customers just got error messages and pictures of the dogs of Amazon, along with a message from Amazon that read: “Sorry, we’re experiencing unusually heavy traffic. Please try again in a few seconds. Your items are still waiting in your cart,” or “”Uh-Oh. Something went wrong on our end.”

Prime Day started at 3 p.m. ET, and the problems emerged almost immediately after.  Around 5 p.m., Amazon tweeted acknowledgement of the problem stating: “Some customers are having difficulty shopping and we are working to resolve this issue quickly.  Many are shopping successfully – in the first hour of Prime day in the US, customers have ordered more items compared to the first hour last year.”

The website Downdetective.com reported that most reported problems came from Amazon’s Website (39%), Log-in (34%)  and Check-out (25%) activities. The site also shows most problems were resolved by 5 or 6 p.m. ET. 

So far, Amazon has not said what the cause of the outage was – some posted about a possible DDOS attack  – but it seems the company’s web servers just weren’t up to the task, despite the company having a year between Prime Days to build up resources

Network-analysis provider ThousandEyes said it detected no ISP outages or content-delivery-network provider errors that might explain the problem, and it also ruled out an overwhelming DDoS attack. “What we did see were application layer errors, which tells us that this was most likely an issue in the web server or a backend API call,” according to a statement by a ThousandEyes’ product manager, Archana Kesavan. “This appears to be an internal Amazon application issue.”

.

One Twitter poster said:

“What everyone’s not realizing about the @amazon outage is that assuming they have 10 million customers browsing at the same time and an average page size of 5MB, the total bandwidth used is 50,000GB/s, or 50TB/s (aka 1,000x Blu-Ray movies-worth each second).”

Prime Day is a big deal for Amazon with some predicting the online giant will make about $4 billion this year on this day alone. Last year the company made between $1 billion to $3 billion or so depending on whose research you read.