Skip to main content

Quick Start

Connect your first cluster and see cost optimization recommendations in under 5 minutes.

Prerequisites​

  • A Kubernetes cluster (any provider or on-premise)
  • kubectl configured with access to the cluster
  • A K8Cost account (free tier works)

Step 1: Create an API Token​

After signing in, navigate to Settings > API Tokens and create a new token. Copy the token value -- you will need it in the next step.

The token authenticates the agent running inside your cluster with the K8Cost API. It is sent as an X-API-Key header and never stored on disk inside the cluster.

Step 2: Deploy the Agent​

Apply the RBAC resources and CronJob to your cluster:

# Create a namespace for K8Cost
kubectl create namespace k8cost

# Apply RBAC (read-only ClusterRole)
kubectl apply -f https://k8cost.com/install/rbac.yaml

# Create the secret with your API token
kubectl create secret generic k8cost-secrets \
  --namespace k8cost \
  --from-literal=api-key=YOUR_API_TOKEN

# Deploy the CronJob agent
kubectl apply -f https://k8cost.com/install/cronjob.yaml

The CronJob runs hourly by default. To trigger an immediate analysis, create a one-off Job:

kubectl create job k8cost-initial --from=cronjob/k8cost-analyzer -n k8cost

Step 3: View Your Dashboard​

Once the agent completes its first run (typically 1-3 minutes depending on cluster size), your data appears in the K8Cost dashboard at k8cost.com/dashboard.

You will see:

  • Monthly cost estimate based on your cluster's resource allocation and cloud-provider pricing
  • Savings potential -- the total monthly savings from all actionable recommendations
  • Top recommendations ranked by estimated impact, with one-click patch generation
  • Cluster health overview showing node utilization, namespace breakdown, and waste categories

What Happens Next​

The agent runs on a configurable schedule (default: hourly). Each run collects current resource usage, evaluates it against 65+ optimization rules, and pushes the results to K8Cost. Your dashboard updates automatically.

As your cluster changes -- new deployments, scaling events, configuration updates -- K8Cost continuously re-evaluates and adjusts recommendations. Stale recommendations are automatically resolved when the underlying issue is fixed.

Next Steps​