Cost Visibility & Allocation
4 entriesSee where the money goes.
Rightsizing & Optimisation
5 entriesMatch requests and nodes to real usage.
Autoscaling
3 entriesAdd and remove capacity automatically.
GPU & AI Workload Costs
21 entriesGPUs can be a large share of an AI cluster bill. These resources cover sharing and scheduling GPUs, measuring utilization, serving models efficiently and tracking LLM spend.
More Cost Visibility Tools & Frameworks
1 entryOpen cost tooling and FinOps guidance.
Related Pages
5 entriesRelated directories on this site.
Frequently Asked Questions
Why is Kubernetes cost hard to see?
Many teams share nodes, so cloud bills must be allocated to namespaces, workloads and teams, and over-provisioned requests hide waste.
What is rightsizing?
Adjusting CPU and memory requests and limits to match actual use, manually with tools such as Goldilocks or automatically with products such as Cast AI and ScaleOps.
OpenCost or Kubecost?
OpenCost is the vendor-neutral CNCF project. Kubecost, now from IBM Apptio, grew out of it and adds enterprise features.
Which Kubernetes cost tools had recent releases?
According to GitHub in October 2026: Karpenter v1.15.1 (23 October), Cluster Autoscaler 1.37.0 (21 October), KAI Scheduler v0.18.6 (15 October), KEDA v2.21.0 (23 September), OpenCost v1.121.3 (16 September) and Goldilocks v4.16.2 (15 September) have recent releases. Kubecost is listed under IBM as IBM Kubecost.
Are the links checked?
Links were checked in November 2026. ScaleOps returned 403 to automated link checks, so its link may need a manual click. Vendor descriptions summarize each vendor's own materials rather than independent evaluations.