infra-kubernetes-cost-optimizer

Detect idle Kubernetes workloads and right-size resource requests with VPA recommendations.

14|1|Updated May 5, 2026
One-click install
npx skills add https://github.com/ivanshamaev/de-agent-skills --skill infra-kubernetes-cost-optimizer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: infra-kubernetes-cost-optimizer
Source: https://github.com/ivanshamaev/de-agent-skills/tree/main/group_skills/infra_dataops_group_skills/infra_kubernetes_cost_optimizer
Command: npx skills add https://github.com/ivanshamaev/de-agent-skills --skill infra-kubernetes-cost-optimizer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps reduce Kubernetes infrastructure cost when cluster bills rise without matching workload growth, mainly by finding idle resources and correcting overly conservative resource requests.

Core Features & Use Cases

  • Detect idle and wasted capacity: identifies zero-replica deployments, dormant workloads, and unused PVC/configurations that still consume budget.
  • Quantify overprovisioning: compares declared CPU/memory requests against actual usage to reveal chronic over-requesting.
  • Right-size safely with VPA/Goldilocks: generates data-driven recommendations for better request values, enabling efficient node packing and lower spend.
  • Implement FinOps chargeback: supports namespace-level cost attribution using tools like OpenCost/Kubecost.
  • Use cost controls: sets LimitRange and ResourceQuota defaults, and applies scheduled scale-down and spot strategies for appropriate workloads.

Quick Start

Install and run the optimizer by enabling Goldilocks in the target namespace (recommendation mode) and then review right-sized CPU/memory requests to reduce idle and overprovisioned capacity.

Frequently Asked Questions about infra-kubernetes-cost-optimizer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce Kubernetes costs when my cluster bill is rising but usage stays flat?

To reduce Kubernetes costs with flat usage, you can detect idle workloads like zero-replica deployments and correct overly conservative resource requests by comparing declared CPU and memory against actual consumption.

What is the best way to right-size Kubernetes resource requests safely?

The best way to right-size Kubernetes resource requests safely is by using VPA and Goldilocks recommendations to generate data-driven CPU and memory values, enabling efficient node packing without disrupting workloads.

How do I set up namespace chargeback for Kubernetes environments?

You can set up namespace chargeback for Kubernetes environments by applying cost attribution tools like OpenCost or Kubecost to allocate shared cluster expenses accurately to specific teams and environments.

Can I use VPA and Goldilocks to fix Kubernetes overprovisioning?

Yes, you can use VPA and Goldilocks to fix Kubernetes overprovisioning by enabling recommendation mode in target namespaces to review and apply optimized CPU and memory requests based on actual metrics.

What Kubernetes controls help limit wasted compute spend across teams?

To limit wasted compute spend across teams, apply Kubernetes controls like LimitRange and ResourceQuota defaults, and use scheduled scale-down and spot strategies for appropriate workloads to enforce budget boundaries.

Does Kubernetes cost optimization work for finding dormant workloads and unused PVCs?

Yes, Kubernetes cost optimization works for finding dormant workloads and unused PVC configurations that still consume budget, allowing you to identify and remove zero-replica deployments and reclaim wasted capacity.