Finout Gets a Handle on Kubernetes Costs

As Kubernetes expands out from the domain of super-users (the internet-scale companies) and towards the enterprise, the cost of running Kubernetes clusters becomes an increasingly important factor in many operations. But how do you measure K8s usage on a per-dollar basis?
Finout has expanded its cost analysis platform for enterprise software to Kubernetes, providing a way to understand the costs of running the open source orchestration tool.
The company’s “cost governance suite” has been updated for Kubernetes so it can be used by managers to manage all the fiscal aspects of running Kubernetes, such as monitoring cloud waste, forecasting future costs and, most importantly, staying on budget.
The platform does not require agents to be attached to the resources, and works with Amazon Web Services‘ Amazon Elastic Kubernetes Service (EKS), Azure Kubernetes Service (AKS), Google Kubernetes Engine (GKE), and others.
“Customers can right-size their Kubernetes workloads in order to be as efficient as they can and actually save money,” said Roi Rav-Hon, CEO and co-founder of Finout. More than just saving money, the Finout information can pave the way for an organization start a “FinOps” (financial operations) practice, where costs can be compared across an entire organization.
FinOut is not the only company trying to get a handle on Kubernetes costs, which promises to be a hot topic at the Cloud Native Computing Foundation‘s KubeCon+CloudNagtiveCon EU this week in Amsterdam.
Last year, another K8s-focused FinOps company, KubeCost, launched the OpenCost open source project, which aims to standardize the reporting around cost tracking, allocation and methodologies. Microsoft Azure is expected to announce at KubeCon today that it is increasing its commitment to this project, following AWS’ lead of doing the same last August.
The Kubecost cost model has been downloaded 5.3 million times between August 2022 and April 14, 2023, according to AWS.
@LiquidReply‘s Vanessa Kantner and Manuela Latz: Over a third of cloud spending is wasted for ever $1 million spent. @rejektsio #Amsterdam @thenewstack #KubeConEU #kubeconeurope pic.twitter.com/bUa3oVabeD
— BC Gain (@bcamerongain) April 17, 2023
More Nuance in the Numbers
The Finout dashboard unifies the organization’s total spend on Kubernetes, even across multiple cloud providers. There is no limit to the number of Kubernetes cluster the customer can fiscally monitor.
The service does not require agents or configuration changes to Kubernetes itself. Instead, it can harness existing monitoring installations, such as Datadog or Prometheus.
With this information on hand, users can track spending against a budget, and identify clusters that may be costing more than anticipated. The platform also includes a feature, called CostGuard, which offers insights into potential ways to reduce usage of Kubernetes.
The offering expands the core service, which tracks other resource-intensive software services, such as Snowflake, Databricks, Datadog and others. Additional integrations will be coming in the future, the company promises.
The service is built from two sets of data. First, Finout parses the customer’s cloud bills, including any discounts, spot-pricing usage, and other factors. It then matches those numbers with the actual usage of the cluster, as captured by Prometheus or some other observability platform. CPU and memory usage along with the configuration information is captured for each pod.
From there, Finout can calculate how much each workload costs, and how much leftover capacity there is on each node. Costs by teams can also be calculated, as can costs-per transaction, per user, per event or other business’ metrics. Customers can also add their own virtual tagging for further nuance.
In addition to the dashboard, all the data collected by Finout is programmatically accessible via an API so that the data can be used and enriched in other ways. One of the options Finout is investigating is around the possibility of an application that automatically allocates workloads according to best cloud costs, for instance.
Check back often this week for all things KubeCon+CloudNativeCon Europe 2023. The New Stack will be your eyes and ears on the ground in Amsterdam!