CLI Installation Guide

This guide shows how to install the Kepler Operator and create PowerMonitor instances using the OpenShift command-line interface. This approach is ideal for automation, scripting, and users who prefer terminal-based workflows.

Prerequisites

Ensure you have:

oc CLI tool installed and configured
Cluster admin privileges on your OpenShift cluster
Current context pointing to target cluster (oc cluster-info)

Step 1: Install the Operator

Create Operator Subscription

Install the Kepler Operator from the community catalog:

oc apply -f - << EOF
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
  name: kepler-operator
  namespace: openshift-operators
spec:
  channel: alpha
  name: kepler-operator
  source: community-operators
  sourceNamespace: openshift-marketplace
EOF

Verify Installation

Wait for the operator to be ready:

# Check subscription status
oc get subscription kepler-operator -n openshift-operators

# Check ClusterServiceVersion (CSV)
oc get csv -n openshift-operators | grep kepler

# Wait for operator pod to be running
oc get pods -n openshift-operators | grep kepler-operator

Expected output when ready:

kepler-operator-controller-manager-xxx   2/2     Running   0   2m

Step 2: Create PowerMonitor Instance

Basic PowerMonitor Configuration

Create a PowerMonitor with default settings:

oc apply -f - << EOF
apiVersion: kepler.system.sustainable.computing.io/v1alpha1
kind: PowerMonitor
metadata:
  name: power-monitor
  labels:
    app.kubernetes.io/name: powermonitor
    app.kubernetes.io/instance: powermonitor
    app.kubernetes.io/part-of: kepler-operator
spec:
  kepler:
    config:
      logLevel: info
      metricLevels:
      - node
      - pod
      - vm
      sampleRate: 5s
      staleness: 500ms
      maxTerminated: 500
    deployment:
      security:
        mode: none
EOF

Verify PowerMonitor Deployment

Check that the PowerMonitor was created and is operational:

# Check PowerMonitor status
oc get powermonitor power-monitor -o wide

# Check conditions
oc describe powermonitor power-monitor

# Verify DaemonSet creation
oc get daemonset -n power-monitor

# Check pods on all nodes
oc get pods -n power-monitor -o wide

Step 3: Verification Commands

Check Operator Health

# Operator pod status
oc get pods -n openshift-operators | grep kepler-operator

# Operator logs
oc logs -n openshift-operators deployment/kepler-operator-controller-manager

Check PowerMonitor Resources

# PowerMonitor status and conditions
oc get powermonitor power-monitor -o yaml

# DaemonSet details
oc describe daemonset -n power-monitor

# Pod logs
oc logs -n power-monitor -l app.kubernetes.io/name=kepler-exporter

Test Metrics Endpoint

# Port forward to metrics endpoint
oc port-forward -n power-monitor svc/kepler-exporter 9102:9102 &

# Test metrics (in another terminal)
curl http://localhost:9102/metrics | head -20

# Stop port forwarding
pkill -f "port-forward.*9102"

Automation Scripts

Complete Installation Script

Save this as install-kepler-operator.sh:

#!/bin/bash
set -e

echo "Installing Kepler Operator..."

# Install operator
oc apply -f - << EOF
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
  name: kepler-operator
  namespace: openshift-operators
spec:
  channel: alpha
  name: kepler-operator
  source: community-operators
  sourceNamespace: openshift-marketplace
EOF

echo "Waiting for operator to be ready..."
while ! oc get csv -n openshift-operators | grep -q kepler.*Succeeded; do
  echo "Waiting for CSV to be ready..."
  sleep 10
done

echo "Creating PowerMonitor instance..."
oc apply -f - << EOF
apiVersion: kepler.system.sustainable.computing.io/v1alpha1
kind: PowerMonitor
metadata:
  name: power-monitor
  labels:
    app.kubernetes.io/name: powermonitor
    app.kubernetes.io/instance: powermonitor
    app.kubernetes.io/part-of: kepler-operator
spec:
  kepler:
    config:
      logLevel: info
      metricLevels:
      - node
      - pod
      - vm
      sampleRate: 5s
    deployment:
      security:
        mode: none
EOF

echo "Waiting for DaemonSet to be ready..."
while ! oc get daemonset -n power-monitor 2>/dev/null | grep -q kepler; do
  echo "Waiting for DaemonSet creation..."
  sleep 5
done

echo "Installation complete!"
echo "Check status with: oc get powermonitor power-monitor"

Make it executable and run:

chmod +x install-kepler-operator.sh
./install-kepler-operator.sh

Uninstall Script

Save this as uninstall-kepler-operator.sh:

#!/bin/bash
set -e

echo "Removing PowerMonitor instance..."
oc delete powermonitor power-monitor --ignore-not-found=true

echo "Waiting for resources cleanup..."
sleep 10

echo "Removing operator subscription..."
oc delete subscription kepler-operator -n openshift-operators --ignore-not-found=true

echo "Removing CSV..."
oc delete csv -n openshift-operators $(oc get csv -n openshift-operators -o name | grep kepler) --ignore-not-found=true

echo "Uninstall complete!"

Configuration Files

For production deployments, save configurations to files:

powermonitor-basic.yaml

apiVersion: kepler.system.sustainable.computing.io/v1alpha1
kind: PowerMonitor
metadata:
  name: power-monitor
spec:
  kepler:
    config:
      logLevel: info
      metricLevels: [node, pod, vm]
      sampleRate: 5s
    deployment:
      security:
        mode: none

powermonitor-production.yaml

apiVersion: kepler.system.sustainable.computing.io/v1alpha1
kind: PowerMonitor
metadata:
  name: power-monitor
spec:
  kepler:
    config:
      logLevel: warn
      metricLevels: [node, pod]
      sampleRate: 10s
      staleness: 1s
      maxTerminated: 1000
    deployment:
      security:
        mode: rbac
      nodeSelector:
        kubernetes.io/os: linux
      tolerations:
      - key: node-role.kubernetes.io/master
        operator: Exists
        effect: NoSchedule

Apply with:

oc apply -f powermonitor-production.yaml

Next Steps

Configuration Guide - Advanced configuration options
Monitoring Setup - Set up monitoring and troubleshooting