Kubecost: How to effectively report Kubernetes container costs

A holistic approach to automation and the necessity of remote project management are just some of the reasons why the demand for containerized applications continues to grow, according to Gartner. Manually monitoring Kubernetes infrastructure costs can be quite a challenge. One solution in this case is to install an automated expense management system, which will reduce the time and effort required for individual container calculations, as well as minimize the risk of errors. An example of such a tool is Kubecost (open source and premium).

What is Kubecost used for?

Kubecost is a real-time data reporting system for teams using the Kubernetes platform. It provides constant visibility into costs, thereby helping to reduce Cloud spending. It should be noted, however, that Kubecost is designed specifically to work with Kubernetes. Therefore, it is not intended for calculating the total costs of deployments involving cloud technologies, which are often used, for example, in RPA implementation. In such cases, appropriate calculation methods should be used, i.e., TCO and TCA. Kubecost uses specific metrics and cost categories that allow for determining, among other things, the amount of consumed and reserved resources (including deployment resources) and their cost, as well as the monthly cluster fee. This type of calculation, in turn, allows for the preparation of recommendations for reducing the aforementioned expenses for Microsoft Azure, AWS, or Google Cloud Platform.

How to manage costs with Kubecost?

The allocation model used by Kubecost allows for cost analysis at any Kubernetes level, supporting its native objects such as namespaces, containers, volumes, or clusters. The data obtained in this way is used, for example, for node optimization. If it turns out that some of them are unused, they can be managed more effectively or simply eliminated. The same applies to storage. One of the newer Kubecost features also helps find the optimal set of nodes by providing recommendations for Kubernetes cluster configuration. This can help users achieve savings of up to 60%. Kubecost is also equipped with a set of notifications that are triggered, for example, when a pre-set cost threshold is reached. Periodic report updates allow for setting realistic financial goals.    

Why implement Kubecost? Pros and cons of using Kubecost

Adding Kubecost to your Kubernetes toolset can help reduce costs and increase company productivity. Detailed analyses allow for the adjustment of operations and protect against the waste of resources. Other benefits include:  

• Real-time cost allocation, configurable via labels. This allows for measuring expenses associated with products and teams. This is possible thanks to the Kubecost Allocation API.
• Dynamic resource pricing after integrating Kubecost with AWS or Google Cloud Platform.
• Cost allocation metrics for CPU, GPU, and storage.
• Alerts and periodic reports.
• Long-term access to metrics.

However, Kubecost does not have automatic scaling mechanisms (e.g., for nodes) that competitors offer. Moreover, this tool focuses primarily on cost optimization. It will therefore likely be a better solution for larger companies whose expenses cover many business units, and it will be less effective for start-ups.

How to implement Kubecost? Kubecost & GitHub

Kubecost offers a free and a premium version. The former is available via GitHub. The open-source Kubecost is, of course, equipped with significantly fewer features, and storage is limited to fifteen days in this case. However, it can be deployed directly in the cluster, run independently, and modified as desired. For a skilled hand, the whole process will be no more difficult than typing a magic phrase like “Kubecost: create by Google” into Google. The monthly cost of the paid version depends on the number of available nodes. As part of the business plan, Kubecost provides multi-cluster support and access to technical assistance. For enterprises with more than two hundred nodes, it is possible to choose the Enterprise plan, which offers unlimited storage and dedicated support.