kubmin is a Kubernetes cost optimization and waste detection platform. It uses energy-powered analysis to find idle workloads, overprovisioned resources, and wrong instance types that are burning money in your clusters.

Beyond cost savings, kubmin also tracks SCI (Software Carbon Intensity) scores for sustainability reporting, giving you both financial and environmental visibility into your infrastructure.

It depends on your cluster size and current waste levels. Typical savings range from $50-500/month per cluster. kubmin's waste detection finds three types of waste:

• Idle workloads that are running but not producing value
• Overprovisioned resources using far less than what they've requested
• Wrong instance types that could run the same workload for less

kubmin takes a fundamentally different approach by using energy consumption data (via Kepler) alongside traditional CPU and memory metrics. This gives kubmin unique capabilities:

• Energy-based idle detection that catches workloads traditional monitoring misses
• Deployment impact tracking to compare image versions across energy, cost, and resources
• DRF-based instance optimization with top 3 alternative recommendations
• SCI sustainability scoring built on the Green Software Foundation specification

We charge a fixed monthly fee with no hidden costs or surprise bills. The free Explore tier gives you basic SCI tracking for 1 cluster.

• Affordable ($19.99/mo) and Premium ($49.99/mo): Both include the complete workload view — waste analysis, instance optimization, deployment comparison, Quick Wins, regional cost comparison, and energy grading
• Premium adds more clusters, workloads, team members, and priority support
• Built-in Prometheus and Grafana included at no extra cost

No. kubmin is completely non-invasive. Just label your workloads and kubmin handles the rest. It works with your existing Prometheus setup or deploys its own stack automatically.

kubmin connects to any Kubernetes cluster — EKS, AKS, GKE, or self-managed. You don't need to change your CI/CD pipeline, deployment tools, or infrastructure. Analysis runs alongside your existing workloads.

kubmin uses Kepler (Kubernetes Efficient Power Level Exporter) for energy measurements. Kepler uses ML-based estimation models trained on real hardware data, which provides excellent accuracy for trend analysis and relative comparison between workloads.

For waste detection, kubmin combines energy data with actual resource utilization metrics (CPU, memory, network), making the waste analysis highly reliable even when individual energy readings are estimates. The pattern detection matters more than absolute numbers.