The very first version of Kubernetes launched in June 2015, and the open-source tool quickly became as essential to containers as containers are to cloud. Today, about 75% of organizations use Kubernetes in production, and adoption continues to rise.

But life with Kubernetes isn’t all hearts and flowers. Sixty-eight percent of users in the 2021 Cloud Native Computing Foundation and the FinOps Foundation survey “Insufficient – or Nonexistent – Kubernetes Cost Monitoring Is Causing Overspend,” reported that their Kubernetes costs are increasing. The problem, as the report title suggests, is an inability to monitor costs.

“About two-thirds of teams can’t answer where they are spending money … and when you can’t answer that question, it’s really hard to be efficient,” said Webb Brown (pictured), chief executive officer at Kubecost (from Stackwatch Inc.), an open-source tool that gives developers visibility into Kubernetes spend.

Brown spoke with David Nicholson, host of theCUBE, SiliconANGLE Media’s livestreaming studio, during a pre-event CUBE Conversation prior to next week’s AWS Startup Showcase: Open Cloud Innovations event. They discussed the importance of Kubernetes cost management. (* Disclosure below.)

Kubernetes adoption set to continue upward trajectory

Kubecost started in 2019 as an open-source project created by a group of engineers working on infrastructure monitoring solutions at Google. This became the basis of Stackwatch’s commercially available Kubecost, which launched in March 2021. Today, Kubecost optimizes resources for over a thousand teams managing more than a billion dollars in Kubernetes-related spend.

“At the end of the day, we like to see teams being more comfortable running their workloads in Kubernetes,” Brown said. “That is the ultimate sign of success.”

Engineers have been forced to pay attention to costs because continuous integration and continuous development cycles make it impractical to wait for the finance department to conduct cost analyses. This has created a “shift-left from a cost perspective” as more and more engineering teams are monitoring costs in real time, according to Brown.

The abstractions and complexities that come with Kubernetes are the major factors driving demand for Kubecost, according to Brown.

“With Kubernetes, you can at any given moment have 50 applications running on a single node or a single [virtual machine]. You can fast forward five minutes and there could be 50 entirely new applications,” he said.

This means that simply assigning the VM or tagging it back to an application or team or department is no longer relevant in most situations, which makes tracking costs extremely complex.

More complexity comes thanks to Kubernetes’ ability to allow distributed engineering teams to work on many microservices concurrently. This is an essential function of the tool but means that there can be no centralized single point of decision-making, as decisions are distributed across not only the infrastructure team, but the engineering team as well, Brown explained.

The third factor that has influenced the need for Kubecost is scale. According to the “2021 Kubernetes Adoption” survey, 89% of respondents expect Kubernetes to play a larger role in the management of their infrastructure over the next two to three years.

“An entirely new ecosystem is being built around Kubernetes,” Mark Potts, managing director and global lead for Accenture’s IBM and Red Hat alliance, told theCUBE last summer. “It won’t be ‘born on the cloud.’ It will be ‘born on Kubernetes and run anywhere’ in the future.”

As adoption of Kubernetes continues to rise, the need for cost management will become more of a business-critical issue.

“Everywhere is where I see Kubernetes in 2022, like very deep in the large and really complex enterprises,” Brown said.

How to know if your company needs container cost management

To find out if companies need container cost management, Brown asks prospective clients what their cost efficiency is for all of their clusters. If they can’t answer that question with a high degree of accuracy — and the majority can’t — then cost management measures are necessary. “We regularly see teams start in the low 20% cost efficiency, meaning that approximately 80% is ‘waste,’” Brown said.

Kubecost tracks costs in any environment where you can run Kubernetes, including on-premises, cloud and multicloud environments. “We give you visibility into all the resources that your Kubernetes clusters are consuming,” Brown added.

Along with direct resources such as nodes, or load balancers, Kubecost can pull in external costs such as Kubernetes tenants that are using Amazon Simple Storage Service (or S3), Google Cloud SQL, or other external cloud services, as well as shared costs. Sometimes this can even include the cost of a DevOps team.

“We’d give you the ability to allocate that back to your core infrastructure, which may be used for show-back or even chargeback across your organization,” Brown said.

Kubecost’s superpower is its foundational layer of visibility that enables the solution to give visibility at any dimension in real-time with a very high degree of accuracy at the largest Kubernetes clusters in the world, according to Brown. That gives users the ability to have real-time monitoring and insights, set alerts and recurring reports, and automate input or data feed decisions.

“All of these things are made possible because of the really hard work that we’ve done to give this real-time visibility with a high degree of accuracy at crazy scale,” he said.

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s pre-event coverage of the AWS Startup Showcase: Open Cloud Innovations event. (* Disclosure: Kubecost sponsored this segment of theCUBE. Neither Kubecost nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE