Stop Paying for Resources You Aren't Using

This one time, at basketball camp, the director told us about the lunch buffet and their one simple rule:

"Take all you can eat. But eat all that you take."

This same rule should be applied to anyone requesting cloud resources. You want a VM with 8GB of RAM? Fine, you can have it, but you had better make use of all 8GB!

In the world of virtualization, as well as cloud computing, the standard is to request more than you will need, just in case it might be needed later. This over-provisioning leads to a lot of wasted resources as well as money. But when you are hosting everything else yourself those costs don't appear as readily as when you get your cloud bill.

Face it, we have no idea how to properly right-size resources when provisioned. And this isn't just about servers, but all manner of resources such as Kubernetes, containers, lambda functions, etc. The spread and sprawl of these services leads to a lot of cost optimization opportunities. So many opportunities we now have tools and consultant services to help optimize cloud spending.

Take for example Platform9 and their Elastic Machine Pool (EMP) product. EMP was purpose built to help identify wasted resources as a result of poorly configured Kubernetes environments. Here is an example of what a common deployment scenario looks like:

In order to stop the madness, EMP provides a simple way to identify inefficient Kubernetes deployments and help right-size them, reduce costs by up to 50% and improving resource utilization by up to 70%.

Once EMP has helped make your deployments efficient, it is likely you will start having your containers and pods hitting resource limits. EMP will also assist by live migrating nodes to other bare metal instances to free up space:

Look, provisioning and managing cloud resources is more art than science. And we make it harder on ourselves by grabbing more at the buffet than we need. EMP helps reduce waste over time and is worth a look if you are reading your cloud bill and wondering why things cost so much. When it comes to optimizing cloud costs, starting with Kubernetes deployments is as good a place as any.

You can read more about EMP here: https://platform9.com/blog/platform9s-elastic-machine-pool-the-most-cost-effective-compute-engine-for-eks/

#CFD19