Kubecost Embraces GCP to Control Kubernetes Costs
At the Google Cloud Next conference last week, Kubecost revealed it made available its platform for managing the cost of Kubernetes clusters on the Google Cloud Platform (GCP).
IT teams that might not be comfortable deploying Kubecost Cloud on Amazon Web Services (AWS) now have another option to access the software-as-a-service (SaaS) application.
Kubecost CEO Webb Brown said as more organizations deploy Kubernetes clusters in production environments, they are incurring cloud costs that are proving increasingly difficult to control and predict. The Kubecost Cloud platform provides those IT teams with both visibility into Kubernetes cloud use along with the controls needed to enforce policies, he added.
In addition, the platform also surfaces recommendations that would enable IT teams to reduce those costs, noted Brown.
At the same time, Google also identified four golden signals for monitoring Kubernetes costs, including workload right-sizing, demand-based downscaling, cluster bin packing and cloud discount coverage. In fact, Google identified right-sizing as the golden signal at the heart of any cost optimization journey; if requests more closely reflect reality, then the decisions Kubernetes automatically makes via setting requests will be more effective.
In general, the level of Kubernetes management maturity has significantly increased as IT teams gain more experience with the platform, said Brown. The challenge is that many IT teams are still overprovisioning Kubernetes infrastructure, which results in developers overprovisioning IT infrastructure, he noted.
In the longer term, cost should be just another metric that DevOps teams are tracking as part of DevOps workflows. In the meantime, the current state of the global economy is making organizations more sensitive to IT costs.
The challenge with allocating Kubernetes costs is that the dynamic nature of container application environments requires a tool specifically optimized for collecting data via Kubernetes application programming interfaces (APIs). In contrast, legacy applications are tied to specific virtual machines that are easier to track.
Of course, many enterprise IT organizations take advantage of enterprise contracts that guarantee discounted pricing if they run a certain number of workloads per month, but there is still a need to track costs by individual business units. In the cloud era, developers have not typically been concerned about costs. As centralized IT teams take more control of the cloud, the lack of cost-containment discipline has become apparent. This often creates a crisis that leads more organizations to embrace FinOps best practices that involve putting programmatic cost controls in place.
As a result, finance teams are asking tougher questions about IT spending than they did at the start of the COVID-19 pandemic when the primary focus was shifting as many workloads to the cloud as possible. The challenge today is organizations simply lack visibility into cloud costs until a bill is presented at the end of the month. By that time, it’s too late to change the outcome.
