Honeycomb sponsored The New Stack’s coverage of Kubecon+CloudNativeCon North America 2020.
StackWatch has expanded the capabilities of its Kubecost tool beyond the management and monitoring of cost and capacity in Kubernetes environments.
Kubecost, launched last year, has now gained the capacity to give visibility into spending on cloud services outside of Kubernetes clusters, the company announced. With this step, Kubecost gives users a single view into both Kubernetes and “out of cluster” spending on cloud resources.
“Kubecost’s new unified cost view allows users to see the complete picture, combining in-cluster spend on nodes, disks, and more with out-of-cluster spend on external services like AWS S3 and RDS,” said Webb Brown, CEO and co-founder of StackWatch. “For the first time, engineering, finance and leadership can get complete visibility across their infrastructure in less than five minutes.”
Previously, Kubecost had minimal support for tagged external resources, so this is its first set of features where both in-cluster and out-of-cluster resources are integrated with significant detail. It’s a much-requested capability from its customers, and represents a significant improvement for Kubecost in this area, according to Brown. The company released the updates in conjunction with Kubecon+CloudNativeCon North America, being held this week virtually.
This new feature is available at no extra cost, and supports Amazon Web Services and the Google Cloud Platform. By merging real-time Kubernetes cost data with external cloud billing data, Kubecost now allows teams to see costs broken down by Kubernetes and other cloud services without any tagging.
With this expanded scope, teams can drill down into cost and capacity data of Kubernetes cluster assets, such as nodes, disks, and load balancers, and of external cloud resources, such as AWS Lambda and S3 storage buckets. Unified data can also be filtered by team, department, project, and more.
Shining a light on Kubernetes Costs
Brown and StackWatch Chief Technology Officer and co-founder Ajay Tripathy got the idea that eventually became Kubecost while working with Kubernetes at Google. “The core problem that we started seeing was that for many teams operating at scale, Kubernetes costs were largely a black box for their developer group,” Brown said. “They didn’t know what the cost of a microservice was. They didn’t even know the cost of a cluster or a namespace or a deployment. It was just really hard to get that visibility.”
At Google, Tripathy was an infrastructure engineer, while Brown was a product manager, both focused on infrastructure monitoring and observability. They realized the lack of visibility into Kubernetes cost and performance capacity was due to Kubernetes environments’ technical complexity and dynamism, and developers’ decentralized decision-making.
“That makes it difficult for teams to stay on top of and have visibility of where they’re spending resources and what team or application is actually driving spend,” Brown said.
Eventually, Brown and Tripathy created Kubecost as an open source project, and later founded StackWatch to offer a commercial version of the product, which is now used by more than 1,000 teams. Kubecost can be used to monitor on-premises and multi-cloud environments. “We are critically focused on being totally agnostic to the environment where Kubernetes is running,” he said.
The product’s main goal is “developer empowerment,” so it’s aimed at application and infrastructure engineers, but it’s also used by finance and accounting teams, as well as by product and executive leaders, according to Brown. Kubecost’s capabilities stretch across three main areas. The first one is around cost visibility and cost allocation. “That’s usually the starting points for teams because when we start working with them, they typically have little or no visibility into costs,” he said.
The second area is the provision of insights and recommendations to be more efficient with resource consumption. These insights can be accessed via the Kubecost UI, or programmatically via the Kubecost API. Finally, Kubecost helps with governance by monitoring thresholds for budgets and resource efficiency and triggering alerts.
StackWatch has also put a lot of emphasis on integrating Kubecost tightly with many cloud-native tools like Prometheus, AlertManager and Grafana. “For example, teams using AlertManager can now have our metrics there and plug them into their own alerting system. Or if they’re using Grafana, they can bring in cost metrics directly into the Grafana UI,” Brown said.
Looking ahead, StackWatch is planning to double its staff to 15 by 2021’s first quarter, and announce its first round of funding. On the features side, it’s eyeing providing more “assistive intelligence” via smart alerting and automated actions.
“We strongly believe we’ve tapped into something powerful here by empowering developers and giving them data that largely wasn’t available, or it wasn’t transparent before,” Brown said. “We’ve seen how important it is to truly democratize access to this data, because teams of all sizes find this information valuable.”
Amazon Web Services and KubeCon+CloudNativeCon are sponsors of The New Stack.