Skip to main content

K8Cost Documentation

K8Cost is a Kubernetes cost optimization platform that helps engineering teams find and fix cloud waste. Connect your clusters, get actionable recommendations backed by 65+ optimization rules, and generate GitOps-ready patches -- all from a single dashboard.

Why K8Cost?​

Most Kubernetes clusters waste 30-60% of their cloud spend on over-provisioned resources. Teams request generous CPU and memory limits during development, then never revisit them in production. Idle workloads run 24/7. Storage volumes sit unused. HPAs are misconfigured.

K8Cost analyzes your actual resource usage against a curated rules engine and tells you exactly what to change -- with the YAML patches to do it.

  • 65+ optimization rules across 13 rulesets covering CPU, memory, storage, replicas, HPA, security, performance, networking, and more
  • GitOps-ready patches in five formats: YAML, Helm values, Kustomize overlays, JSON Patch, and Strategic Merge Patch
  • Real-time anomaly detection using z-score analysis to catch cost spikes before they hit your bill
  • Multi-cluster visibility with a single dashboard across AWS, GCP, Azure, and on-premise clusters
  • Every feature on every plan -- you only pay for cluster scale, never for features

How It Works​

  1. Deploy the agent -- A lightweight CronJob runs inside your cluster, collecting resource metrics via the Kubernetes API. Read-only access, no cluster-admin required.
  2. Rules engine evaluates -- 65+ rules analyze your pods, nodes, HPAs, PVCs, and namespaces against actual usage data and cloud-specific pricing.
  3. Recommendations appear -- Each recommendation includes severity, estimated savings, and a one-click GitOps patch.
  4. You apply or automate -- Approve recommendations individually, export patches for your CI/CD pipeline, or create PRs directly on GitHub and GitLab.