Skip to main content
Account

OpenCost

Learn how OpenCost attributes Kubernetes spend to namespaces, pods and labels using Prometheus metrics and cloud billing data.

Status
Stable
Category
Observability and Analysis
Subcategory
Continuous Optimization
License
Apache-2.0
Technology matrix Skip Just not for me See the matrix

OpenCost is a cost monitoring tool for Kubernetes that attributes cloud infrastructure spend down to individual namespaces, deployments, pods, and labels. It is the open-source core extracted from Kubecost, donated to the CNCF, and maintained as the reference implementation of the FinOps Foundation’s Kubernetes cost allocation spec.

OpenCost joins Prometheus metrics (kube_pod_container_resource_requests, container_memory_working_set_bytes, etc.) with cloud provider billing data to compute hourly costs for CPU, memory, GPU, storage, and network for every workload. It ships with built-in pricing for AWS, GCP, Azure, Oracle, Alibaba, and Scaleway, and can consume the AWS Cost and Usage Report or GCP/Azure billing exports for reconciled “actual” costs rather than list-price estimates. Allocation output is available via a REST API, exported as Prometheus metrics, and increasingly via the FOCUS billing schema for ingestion into downstream FinOps tools.

The project is Apache-2.0 and a CNCF incubating project. In the same space you will find Kubecost (the commercial superset), Cloudability, CAST AI, Vantage, Finout, and cloud-native tools like AWS Cost Explorer — but OpenCost is the only vendor-neutral option that runs entirely inside the cluster and speaks Prometheus natively.

Continuous Optimization

Email signup loads here. You can also manage email preferences in account settings.

Build OpenCost? I'll show you where engineers get stuck with it → Back free content for the community →