Over 20 years ago, Amazon[1] started its march toward becoming the top retailer, on or offline, in the world. Ten years ago, Amazon Web Services[2] began its journey to becoming the No. 1 public cloud provider anywhere. Today, July 16, 2018, Amazon is blowing years of work and billions in value, as both its sales site and cloud collapse under the weight of Prime Day 2018.
Read also: 5 Prime Day deals you shouldn't miss | Amazon Prime Day 2018: The best early deals in tech so far[3][4]
Overwhelmed by millions of eager shoppers, Amazon and its vaunted services are having perhaps their worst day ever. Instead of great buys, customers are running into errors featuring cute, but not for sale, dogs.
Amazon hasn't said why this is happening. One snarky user[5] suggested that "Amazon.com must not be following the Well-Architected Framework Guide[6]." This is Amazon's own textbook on how to build cloud applications.
Amazon has been running on AWS since 2011. As then-Amazon engineering lead Jon Jenkins explained, "We started thinking in 2009 about how to use it [AWS] for Amazon.com[7]. ...On November 10, 2010 is the day we turned off the last physical web server serving Amazon.com."
Alas, even with years more experience, AWS didn't prove up to the challenge. Amazon still has to fix up its own dog food.
My own informed guess, based on what I've been seeing, is Amazon front-end caching and AWS Elastic Load Balancing (ELB)[8] have proven not up to the job with dealing with millions of simultaneous shoppers.
You can mitigate Amazon's