A software glitch at one of the Amazon, Inc. (NASDAQ:AMZN)’s data centers left a host of high profile web services offline. Users of different services like Instagram, Netflix, Inc. (NASDAQ:NFLX), Vine and Airbnb faced service issues for many hours on Sunday, says a report from BBC. All these services are dependent upon Amazon’s cloud based network.


One of Amazon’s data centers in North Virginia ran into trouble and could not stand up to demand. Cloud softwares like Heroku and web automation service IFTTT also faced similar problems.

Instagram was the first to report issues

The problem was first traced by Instagram when it found that access to its service was intermittent  through Twitter, and other services started showing warning messages within a few minutes.

Users in the United States started facing problems at around 16.00 Eastern Time, which persisted for another several hours. During the failure web services were sporadic, users said that they could not easily access the login pages, and the sites were slow to update.

Reasons for the glitch at Amazon’s data center

Amazon, Inc. (NASDAQ:AMZN) reported that it is trying to unearth the various issues at the Virginia data center. The problem affected databases, software that spreads queries on different servers,and the code controlling the main computers that backed the whole service.

Two reasons were discovered, from the logs detailing the outage, for the failure that continued for five hours. On Sunday at 4:29 PM ET Amazon reported that EBS volumes in its ECS database were experiencing launch and network packet loss errors.

The second problem was fixed in an hour, but some of the sites were still facing the issues, and only a few could be restored before 9 PM ET. Amazon, Inc. (NASDAQ:AMZN) was also facing the glitch with load balancers in the U.S. East 1 region. The problem first came into light at 4:40 PM and was fixed in an hour but was fully repaired in 4.5 hours.

After fixing the problem, Amazon, Inc. (NASDAQ:AMZN) said that it was caused by a “partial failure of a networking device” but gave no further details.

A week before, Amazon’s North American web stores went offline for about 30 minutes.