6 Comments

Summary:

Amazon Web Services experienced a brief outage on Sunday afternoon. It only last about 60 minutes, but appears to have taken down popular sites such as Instagram, Flipboard and Vine for short periods.

aws outage

Amazon Web Services, the world’s biggest and most well-known cloud computing provider, once again had some availability problems in its US-EAST data center.

The outage stemmed from a network issue and lasted from 12:51 p.m. Pacific Time on Sunday until 1:42 p.m., according to the AWS status page. As of 3:23 P.M., AWS was reporting that most affected Elastic Compute Cloud instances were back up and running and that it was “continuing to work on a small number of instances and volumes that require additional maintenance before they return to normal performance.

It’s hard to say how many sites were affected by the brief outage, but it appears Instagram was, at least minimally. One of the company’s engineers tweeted about the problems pretty early on.

Numerous reports on Twitter noted that Flipboard and Vine were down, as well.

Airbnb acknowledged some problems, as well.

2013 has actually been a pretty good year thus far for AWS, with this being the first outage (at least that I recall; please correct me if I’m wrong) and brief one at that. Last year, however, ended on a sour note with a Christmas Eve outage that even took down Netflix — a guiding light for all developers trying to build resilient application atop Amazon’s seven-year-old, but still evolving cloud platform.

Last week, the Amazon.com front page went down for about 45 minutes and cost the company an estimated $5 million.

The US-EAST region is AWS’s biggest, most-used (and least expensive (although its US-WEST-2 region in Oregon now offers the same pricing) region, which means that outages there often affect a large number of web sites and services. Best practices are beginning to emerge (from places other than Netflix even) about how architect applications that survive an outage in one Availability Zone, but not everyone heeds this advice, and sometimes even the best-laid plans aren’t good enough.

Here is the full AWS report on Sunday’s incident:

aws out

 

Update: This post was updated at 8:08 p.m. on Aug. 25 to clarify that Amazon Web Services’ US-WEST-2 region in Oregon now offers the same prices as its US-EAST region.

  1. The Amazon major outages seems quarterly event to me. I did research on Amazon cloud major outages and found some interesting aspects of challenges Amazon has http://www.slideshare.net/rahultyagi50999/amazon-cloud-major-outages-analysis.

    Share
  2. Inderaj (Raj) Bains Monday, August 26, 2013

    AWS outage – once burned, twice shy – ShortStack for Facebook switched from Heroku on AWS to Clustrix on last outage, survived this one.

    http://www.clustrix.com/blog/bid/312144/AWS-outage-once-burned-twice-shy-ShortStack-for-Facebook-switched-from-Heroku-on-AWS-to-Clustrix-on-last-outage-survived-this-one

    Share
  3. Wendy Purcell Tuesday, August 27, 2013

    It’s easy to dodge bullets like this. Go to a provider that actually guarantee’s up-time and has to live by it. Having phone support also helps. I left AWS for Codero Hosting last year and haven’t looked back. http://www.codero.com

    Granted yeah the feature set isn’t completely there but what they miss on crazy EBS/EC2 they make way up for in price and service.

    Share
  4. Yeah,

    That is why OpenStack powered by Rackspace guarantees 99.999% uptime as well as a 1 Hour SLA. People may state that OpenStack used EC2, but what they don’t understand is that it only uses the 3rd party API’s

    Share
    1. Let me restate that…

      That is why OpenStack powered by Rackspace guarantees 99.999% uptime as well as a 1 Hour SLA. People may state that OpenStack uses EC2, but what they don’t understand is that it only uses the 3rd party API’s!!! Plus you a staff that will guide you through any issue you may run into that is bundled with your cloud infrastructure!

      Can EC2 do that????

      Share
  5. For those who know – I am assuming this less than 1% of aws. Grow up and not generalize.

    Share

Comments have been disabled for this post