In recent years, redundancy in network infrastructure has shifted from a commodity to an absolute necessity. You may remember (back before the cloud) businesses scaled their infrastructure vertically. Bundling services from a single provider was a popular (and often more affordable) solution.
With the shift towards cloud migration, vertical scaling is no longer the best option, as it can be more expensive and more commonly leads to service disruptions. Many clients of Amazon Web Service’s S3 service learned this lesson the hard way today.
@awscloud Down and everything comes to a screeching halt. Reminding us that you should never have all your eggs in one basket. #AWSPOCALYPSE
— Sam Clark (@McClark71) February 28, 2017
This afternoon, around 1:30pm EST the world was yet again reminded of the importance of redundancy. Many Amazon S3 clients were unavailable on the East Coast for hours.
According to Amazon’s status site, issues within their S3 service were causing “high error rates”. Trello, Quora, Slack, IsItDown.com, and many other websites struggled to stay online.
The moment the presenter realizes S3 is down. #AWS #S3 #Amazon pic.twitter.com/S8gY1J9lrH
— Ian Sherwood (@ian_surewould) February 28, 2017
President of DNS Made Easy Steven Job commented that,
“In the future, we recommend that clients use DNS Failover with CNAME records to stay online even when their content delivery provider is experiencing issues.”
This functionality is currently available in Constellix DNS. CNAME Failover will automatically detect if the primary provider is downed using Constellix Sonar monitoring checks. If the provider is downed, Constellix DNS will automatically failover traffic to the secondary provider.
