Kubecost

Kubecost is an other FinOps tool that allows you to monitor and optimize the costs of a Kubernetes cluster. once it is deployed. The tool automatically calculates the cost of each object deployed according to the resources used (CPU/GPU/RAM/PV…) and calculates the efficiency of the object in order to better optimize the allocation of resources, thus obtaining the optimal cost.

Here is an overview of what Kubecost offers:

We observe with the overview that it is possible to save 34.75$ from 147.11$ of expected monthly costs, which would give a total of: 147.11 - 34.75 = 112.36$ monthly costs; Kubecost helps to save 24% of the expenses!

To observe the costs of allocated resources and the efficiency of Kubernetes objects, head to the Cost Allocation section:

And to start optimizing your costs with Kubecost, head to Savings:

Here, Kubecost proposes several solutions to save costs: use AWS Reserved Instances, resize the local disks or right-size the container requests.

👉

It is important to remember that these optimizations must be applied with a LOT of attention, for example there is no point in resizing disks dedicated to monitoring if you know that they quickly tend to double or even triple the space used, this kind of detail is not taken into account by Kubecost: it is the role of a FinOps operator to know this.

Let’s take a closer look at how Kubecost helps us to Right-size container requests:

Kubecost, in addition to offering Auto Recommendations for over-provisioned containers that cost more than they consume, also deals with under-provisioned containers that might not perform as well as expected.