Skip to content
Start free

Kubernetes cost monitoring

Know what every Kubernetes workload costs.

Cost per namespace, workload and pod on EKS, GKE, AKS and bare metal, split by team, with rightsizing changes you paste into Git. One Helm chart.

Start freeBook a demo
  • 5-minute install
  • Every permission listed
  • Free up to 100 vCPU

Example: one workload, from cluster to container

Cluster

$34.8K

prod-eu-1

48 nodes

Namespace

$4.8K

data-platform

142 pods

Workload

$2.1K

analytics-worker

12 pods

Pod

$180

analytics-worker-78f9

4 cores

Container

$43

worker

cpu 4× over

Works with
  • AWS
  • Google Cloud
  • Azure
  • OpenShift
  • Bare metal

Network Intelligence · Early access

Kubernetes cost, including the network.

See which workloads send traffic across zones and out to the internet, and what that traffic costs. Early access is by request.

Request early accessRead how it works
  • Traffic between zones and out to the internet, per workload and per namespace.
  • Measured by an optional eBPF agent that runs as a privileged DaemonSet (Linux kernel 5.10 or newer).
  • It records flow byte counts, DNS names, TLS server names and HTTP method, route and status. It does not record request or response bodies.

How it works

Install in 5 minutes.

Costs show up right away. Rightsizing recommendations follow once there is about 7 days of usage history.

  1. helm installprod-eu-1

    Step 01

    Connect

    Add the Helm repo and install the chart with your cluster token. The agent runs in the kubeadapt namespace and sends usage snapshots to Kubeadapt over HTTPS.

  2. 4 of 16 over-provisioned

    Step 02

    See cost

    Costs show up within minutes, broken down by namespace, workload and node. Assign namespaces to teams to get per-team totals.

  3. $34,827−$5,180$$29,647

    Step 03

    Rightsize

    Each recommendation shows the YAML change, the monthly saving, and the usage behind it. Copy it into your repo or dismiss it.

helm repo add kubeadapt https://kubeadapt.github.io/kubeadapt-helmhelm repo updatehelm install kubeadapt kubeadapt/kubeadapt \  --namespace kubeadapt --create-namespace \  --set agent.config.token="<your-token>"

Get the token from app.kubeadapt.io → Clusters → Add cluster. Supports Kubernetes 1.24–1.35.

Product tour

Every Kubernetes object has a cost page.

Workloads, pods, nodes, node groups and namespaces each show cost over time, usage against requests, and open recommendations.

Example · Workload

Cost and usage per workload.

Open any Deployment, StatefulSet, DaemonSet or Job to see what it costs, how much of its requests it uses, and which containers ask for more than they need.

  • Cost trend & breakdown

    Run-rate over time, split into CPU and memory.

  • CPU + memory vs requests

    Usage curve with requested dashed reference and per-pod toggle.

  • Hourly utilization heatmap

    Peak-hour patterns to inform HPA min/max.

Cost Explorer

Group and filter cost by any label.

Namespace, team, label, region, instance type or workload. Save views as dashboards and share them.

  • Filter
    Cluster, namespace, label, team, department, instance type, zone, region, node group, spot or on-demand.
  • Group
    By namespace, workload, controller kind, node, pod, container or label.
  • Saved dashboards
    Save a view as a dashboard and share it with your team.
  • Export
    Download CSV or PDF, or pull the FOCUS-format export into your warehouse.
Example data: one cluster’s cost for 30 days, grouped by namespace.

What you get today

Cost monitoring, cost by team, rightsizing, recommendations, alerts, configuration checks and carbon estimates.

On the roadmap

  • SecureOpsSecurity posture checks
  • GPU & LLM OptimizationGPU and LLM cost
  • Shift-Left Cost PreventionCost checks before deploy

Install in 5 minutes

Stop paying for unused capacity.

Free up to 100 vCPU. After that, $1.99 per vCPU per month, dropping to $0.60 at volume.

Start freeBook a demo
Agent runs in your clusterNo credit card