Installation¶

Complete Pre-requisites first (Kubernetes version, Prometheus, PostgreSQL, and tooling). The Helm command below assumes you already have a compatible Prometheus URL — see Prometheus scenarios if you are unsure which setup to use.

With Helm¶

Install from the OCI registry (replace the Prometheus URL with yours):

helm install cruisekube oci://tfy.jfrog.io/tfy-helm/cruisekube \
  --namespace cruisekube-system  \
  --create-namespace \
  --set cruisekubeController.env.CRUISEKUBE_DEPENDENCIES_INCLUSTER_PROMETHEUSURL="http://prometheus-kube-prometheus-prometheus.monitoring.svc:9090"

Set and verify the Prometheus URL

Replace the example CRUISEKUBE_DEPENDENCIES_INCLUSTER_PROMETHEUSURL with the in-cluster Service URL of the Prometheus instance CruiseKube should query (not localhost). Confirm the Service exists and is reachable from the controller namespace before you install:

kubectl get svc -A | grep -i prometheus

This URL directly affects whether CruiseKube can fetch metrics — wrong or unreachable endpoints lead to empty or incomplete recommendations. See Prometheus prerequisites to pick the right instance, and Troubleshooting — Prometheus metrics if metrics look missing after install.

Customize any installation with a values.yaml file.

Uninstall¶

helm uninstall cruisekube -n cruisekube-system
kubectl delete namespace cruisekube-system