Skip to main content

Amazon’s Servers Go Down, and The Internet Comes to a Halt

493800514_8d129be64b_oOn Sunday morning we were reminded that, contrary to current cosmological theory, the world really revolves around an inconspicuous data center in Virginia.

A server fault in an Amazon data center in Virginia earlier this morning took down many Amazon services and many popular websites. Amazon Web Services is the Amazon division responsible for hosting web and mobile apps. It’s distributed across multiple locations, and supports websites all over the world.

And it crashed hard on Sunday. There are numerous reports on Twitter that a surprising number of web services ranging from Netflix to Medium and Nest to Buffer and Pocket all became unavailable on Sunday morning.

The Sunday morning crash also shut down Echo, Amazon Instant Video, the retail functions of, IMDb, Createspace, Mayday, and the Kindle Store were all affected.

Len Edgerly of The Kindle Chronicles reported several of the downed services on Twitter this morning. He said that he couldn’t order a Fire tablet, or use his Fire Stick to stream video (curiously, I had no trouble with Amazon Video during that period).

For the most part only customers in the Northeastern US were affected, but I also have reports from Germany that the Kindle Store on was down, and so was Createspace. Another German Amazon customer reported that her Fire Stick stopped functioning at about the same time.

According to The Independent, other affected services include Tinder, Product Hunt, SocialFlow, GroupMe, Viber, and Mediacom.

Far more services were affected in the Sunday morning crash than the last time AWS crashed, in 2013. Techcrunch reported at the time that the crash took IFTTT, AirBnB, Vine, Netflix, and Instagram offline.

That 2013 crash also happened on a Sunday, and it also involved an Amazon data center in Virginia. That crash lasted less than two hours before most of the services were back online.

Today’s crash ran past the four-hour mark before Amazon’s services were restored, and according to the AWS Health Dashboard the system is operating normally again.

image by versageek


Similar Articles


Rob Siders September 20, 2015 um 5:10 pm

I was unable to use the search function on the Fire Stick, but could watch content fine. I also tried to place an order, but it would fail while trying to verify my address.

DebbyS September 20, 2015 um 6:10 pm

Here in Albuquerque, I had trouble with YouTube, though I could take YT urls I subscribe to (they send me "hey, new content!" emails) and I put them in my YouTube downloader and that worked okay. But I couldn’t go into the site and thumbs up anyone. Twitter and Local Yelp worked okay, but other sites didn’t, so while I worried that my just updated I.E. 11 wasn’t liking some sites, I was also thinking it was maybe some solar influence (there was a minor solar storm effecting Earth) or something along that line. Here it is 4:09pm mountain time and everything is working, whew!

Maria (BearMountainBooks) September 20, 2015 um 7:41 pm

I post bargain books every morning either on my blog or on a FB page for such and as I was doing my posts, there were no actual buy buttons on the book pages at Amazon! (I’m in Texas). By the time I finished with the post, the buy buttons were back. First comment on the book? "There’s no buy buttons." I looked. Yup, gone. I went to do a post on being patient with the buy buttons and…they were back.

The author’s dashboard was also down for a couple of hours and another author reported that she was unable to mail .mobi files via Amazon’s service to kindles.

Tracey September 20, 2015 um 11:03 pm

Still having GroupMe issues, failing to send messages. Many hours after the fact. Located in NJ.

Nate Hoffelder September 21, 2015 um 8:18 am

It’s surprising just how many services depend on Amazon, isn’t it?

DebbyS September 21, 2015 um 9:00 am

From a scientifically oriented YT channel…

Did the Sun affect the San Diego Blackout? S0 News Sept 21, 2015

Maria (BearMountainBooks) September 21, 2015 um 9:53 am

Part of KU reporting is still down/delayed according to the message that flashes up on one area of the dashboard.

Tymber Dalton September 21, 2015 um 10:31 am

And people say Mercury Retrograde is bullshit. 😉

Fortunately, I slept through most of this, except when I got up and tried to access KDP’s backend I couldn’t for about fifteen minutes.

Nate Hoffelder September 21, 2015 um 6:38 pm

@ Tymber I was wondering how far we were from reenacting "The Machine Stopped."

Eight Things to Look for in a Good Web Host – Tauni Everett February 17, 2016 um 4:55 pm

[…] Uptime Guarantee. Uptime is the percentage of time the hosting server has been up. Every host encounters equipment failures, power outages or loss of service which can result in downtime for clients. Just last fall Amazon’s servers were down which caused a ripple affect that cost over $5 million and affected… […]

Write a Comment