Avoiding Simple Code Breaks on AWS

February 15, 2017 General, General

Faulty code remains one of the most common reasons for performance issues and unscheduled downtime. Yet, it is still possible for bad code to make its way into the Production environment, even with testing.

Your Production environment is most likely a collection of interlinked servers, load balancers and more. Does it really make sense to test your code on a single sandbox server that doesn’t equate in any way to the real world it will need to operate within? Of course your code performs flawlessly within the simplified, unchallenging environment of a single testing or staging server. But that’s no indication of what may happen upon release when it attempts to mesh with the harsh realities of your full Production environment.

Instead, code should be tested in a complete and exact replica of the current Production environment, right down to the last load balancer. Only then will you know if there are potential conflicts or if something will break under increased traffic load.

Thing is, it’s actually a doddle to create such replica environments in minutes. Platforms (such as Anchor Fleet) can completely automate the building of predefined infrastructure on demand. And as you can destroy testing environments when not in use, it can cost a lot less than maintaining permanent testing and staging servers.

Find out more about best AWS practices, by downloading our AWS EBook!