How to save money on scaling in the cloud

This is probably a situation every cloud developer can relate to: It’s a slow friday, you just triggered a compilation, and while slacking off a bit, you decide to go ahead and take a look at your environments metrics dashboard and you may see something like this:

It might even be lower! You think to yourself: Awesome, we are handling load so well, we barely ever have to use all the resources available.

Well, that is indeed true, but while having spare resources for your workloads is a great situation to be in, it is also actively costing you or your company a lot of money which is basically thrown out of the window.

Overprovisioning

The situation seen in the screenshot above is also described as “Overprovisioning”. Overprovisioning in regards to the cloud is understood as the situation in which e.g. a virtual machine is allocated for a given application, which provides way too much resources the application would never need or be able to use.

While having too much of anything is usually a great thing and much better than having too few of something, it usually also comes with a cost.