More on historical AWS outages and lessons

[jlevy]

More links like these:
http://royal.pingdom.com/2011/04/27/the-amazon-ec2-outage-time-to-learn-from-it/

[rjhintz]

Just parking this link here, though it's about the 2011 outage, since it and the follow on it references are a good resource. The Big List Of Articles On The Amazon Outage

Another: What we learned from the AWS Outage associated with this video with potato quality audio. Interesting that the speaker mentions aligning the AZs for all his accounts with the same numbering.

Other references, to be reviewed for applicability:

• Paper Review: Why Does the Cloud Stop Computing? Lessons from Hundreds of Service Outages
• A List of Post-mortems!

[danhermann]

It was smaller in scale, but the DNS outage two weeks ago took down a number of AWS services in the us-east-1 region. My apps couldn't access SQS and others reported problems accessing DynamoDB and Kinesis. https://news.ycombinator.com/item?id=12763768

[bwagner5]

Oct 2012 Outage and Netflix's Post Mortem:
http://techblog.netflix.com/2012/10/post-mortem-of-october-222012-aws.html

EBS Outage can cause problems for EBS backed storage. Netflix switched to almost completely instance-store backed EC2.

[jlevy]

With a little creative Googling found a list of major post-mortems:

EC2, EBS, RDS in eu-west-1, 20??-08-07
EBS and EC2 in ap-southeast-2, 20??-06-04
SimpleDB in us-east-1, 20??-06-13

RDS and EC2 in us-east-1, 2011-04-29
One AZ in us-east-1, 2012-07-02
EBS and EC2 in us-east-1, 2012-10-22
ELBs in us-east-1, 2012-12-24
AZ in sa-east-1, 2013-12-20
DynamoDB, SQS, AutoScaling, CloudWatch in us-east-1, 2015-09-20 (article for dating year)

[jlevy]

Latest one: S3 and many other services in us-east-1, 2017-02-28