Copyright SiliconANGLE News
Apptio, the cloud analytics firm that IBM Corp. acquired in 2023, today is launching expanded and enhanced software products aimed at helping companies gain greater control over cloud spending amid accelerating investments in artificial intelligence workloads. The release includes Cloudability Governance, a cost compliance tool integrated with the Terraform infrastructure-as-code manager, as well as version 3.0 of Apptio’s Kubecost Kubernetes cost monitoring platform. Both provide organizations with real-time visibility and control over cloud and container infrastructure costs. “Generative AI is not only pushing the limits of cloud infrastructure; it’s challenging the ability of technology and business leaders to make informed decisions and evaluate tech spending return on investment,” said Eugene Khvostov, Apptio’s chief product officer. Apptio cited its own research that found that 55% of business leaders lack the information needed to evaluate technology spending effectively. Cloudability Governance now integrates with the Cloud Platform from recent IBM acquisition HashiCorp Inc. and Terraform to support cost estimation, policy compliance and automated monitoring at the time of deployment. Khostov said platform engineers can use the data to see the cost impact of their decisions. “We’re able to place a gate or a toll booth that looks at what the engineer is going to deploy in a version control system like GitHub and tell them that here’s what it’s going to cost,” he said. Apptio connects to major cloud providers and uses data provided by customers to generate cost estimates. “We encode what your private rates are, what your custom rates are or might be and calculate costs in real time,” he said. “We bring in trillions of rows of records a month and detect blind spots like legacy instances, over-provisioning and opportunities to purchase committed instances for significant discounts.” Tag management The platform also addresses common issues with tagging cloud resources for ownership and cost allocation. Cloud customers use tags to identify the owners and users of a resource. “If a team is running or initiating a resource and forgot to apply its tag, it creates a blind spot,” Khostov said. Cloudability can identify who initiated a resource and automatically apply tags. The software also supports synthetic tags, which map to the business. Kubecost 3.0 adds container-specific insights to the FinOps toolkit. It runs inside a cluster, gathers real-time metrics and can automatically allocate costs while also providing optimization recommendations. Kubecost is container-agnostic and monitors Kubernetes usage across both on-premises and cloud infrastructure. The new version introduces graphics processing unit-specific metrics through Nvidia Corp.’s Data Center GPU Manager exporter. “You can look at GPU-specific utilization, memory utilization and a host of different metrics,” Khostov said. GPUs are critical to AI training and inference. Kubecost is integrated with Cloudability’s FinOps suite to enable bidirectional sharing. “It’s a single pane of glass for understanding your cloud and Kubernetes spending,” Khostov said. “You can see how costs are broken down by team, but also by factors like container constructs, namespaces and labels.” The recent spike in AI development activity hasn’t created new blind spots but has increased the velocity at which costs grow, Khvostov said. “You might be spending thousands of dollars an hour on GPUs, and if you don’t have visibility into that in real-time, the impact can be significant,” he said. Cloudability Governance is now in public preview. Kubecost 3.0 is generally available. Both will be on display at KubeCon North America next week. Image: Adobe Stock
