While Amazon is not disclosing what exactly caused its site to fail for two hours on Friday, the world's largest online retailer may find its problem comes from the site design itself.
Amazon.com first failed around 1:25 pm EDT on Friday. For a two-hour period, customers attempting to access the site were greeted with an error message. Once the site began to return, portions of Amazon remained largely inaccessible.
Problems seemed to appear once again on Monday, as some users reported receiving "Service Unavailable" errors. This was the same message that appeared on Friday.
The problem began around 1:10 pm EDT Monday, and lasted for about 15 minutes. Site issues seemed to continue through about 3pm.
"Amazon's systems are very complex and, on rare occasions, despite our best efforts, they may experience problems. We work to minimize any disruption and to get the site back as quickly as possible," the company said in a statement Friday on the downtime, declining to comment further.
Amazon Web Services, the company's effort to bring its "cloud computing" power to smaller Internet companies, appears to have remained online during the incident. The Web Services page was down during the period, but customers reported their applications based on AWS did not fail.
Such issues are generally rare for Amazon. The last outage of major consequence was Thanksgiving Day 2006, when a sale on Xbox 360s brought the site down temporarily several times.
Analysts said the most likely reason for the downtime was human error. Keynote Systems, a firm that monitors Web activity, pointed to a configuration file glitch. It is unclear whether the same problem succeeded in bringing down Amazon again.
