Master Cloud‑Native Monitoring: Deploy Prometheus Operator with Helm

What is Prometheus Operator?

Prometheus Operator is a Kubernetes Operator that provides native deployment and management of Prometheus and related monitoring components.

Key Features

Kubernetes custom resources: Deploy and manage Prometheus, Alertmanager, and related components via CRDs.

Simplified deployment configuration: Configure basic Prometheus settings such as version, persistence, retention policy, and replica settings from native Kubernetes resources.

Prometheus target configuration: Automatically generate monitoring target configurations based on familiar Kubernetes labels without learning Prometheus‑specific configuration language.

Deploy Prometheus Operator

1. Download the Helm chart and push it to a Harbor repository:

<code>$ helm repo add prometheus-community https://prometheus-community.github.io/helm-charts --force-update
$ helm pull prometheus-community/kube-prometheus-stack --version 50.3.1
$ helm push kube-prometheus-stack-50.3.1.tgz oci://core.jiaxzeng.com/plugins
Pushed: core.jiaxzeng.com/plugins/kube-prometheus-stack:50.3.1
Digest: sha256:71397b86f144955df242381462e200fdd107de99d57c935cede9b38f67892547</code>

2. Pull the chart on the Kubernetes host:

<code>$ sudo helm pull oci://core.jiaxzeng.com/plugins/kube-prometheus-stack --version 50.3.1 --untar --untardir /etc/kubernetes/addons/</code>

3. Create a custom values file (excerpt):

<code># Custom resource name
fullnameOverride: "monitor"
cleanPrometheusOperatorObjectNames: true

# Custom images
prometheusOperator:
  admissionWebhooks:
    patch:
    image:
      registry: core.jiaxzeng.com
      repository: obs/kube-prometheus-stack/kube-webhook-certgen
prometheusConfigReloader:
  image:
    registry: core.jiaxzeng.com
    repository: obs/kube-prometheus-stack/prometheus-config-reloader
    image:
      registry: core.jiaxzeng.com
      repository: obs/kube-prometheus-stack/prometheus-operator

# kube‑state‑metrics image
kube-state-metrics:
  image:
    registry: core.jiaxzeng.com
    repository: obs/kube-prometheus-stack/kube-state-metrics

# node‑exporter image
prometheus-node-exporter:
  nameOverride: node-exporter
  image:
    registry: core.jiaxzeng.com
    repository: obs/kube-prometheus-stack/node-exporter

# Prometheus image and storage
prometheus:
  prometheusSpec:
    image:
      registry: core.jiaxzeng.com
      repository: obs/kube-prometheus-stack/prometheus
    storage:
      volumeClaimTemplate:
        metadata:
          name: data
        spec:
          storageClassName: longhorn
          accessModes: ["ReadWriteOnce"]
          resources:
            requests:
              storage: 20Gi

# Alertmanager image and storage
alertmanager:
  alertmanagerSpec:
    image:
      registry: core.jiaxzeng.com
      repository: obs/kube-prometheus-stack/alertmanager
    storage:
      volumeClaimTemplate:
        metadata:
          name: data
        spec:
          storageClassName: longhorn
          accessModes: ["ReadWriteOnce"]
          resources:
            requests:
              storage: 10Gi

# Disable Grafana
grafana:
  enabled: false

# Enable collection ports for etcd, controller‑manager, scheduler
kubeEtcd:
  service:
    enabled: true
    port: 12381
    targetPort: 12381
kubeControllerManager:
  service:
    enabled: true
    port: 20257
    targetPort: 20257
kubeScheduler:
  service:
    enabled: true
    port: 20259
    targetPort: 20259</code>

4. Install the operator with Helm:

<code>$ helm install --create-namespace -n obs-system monitor -f /etc/kubernetes/addons/kube-prometheus-stack-values.yaml /etc/kubernetes/addons/kube-prometheus-stack
NAME: monitor
LAST DEPLOYED: Wed Jun 11 10:50:43 2025
NAMESPACE: obs-system
STATUS: deployed
REVISION: 1</code>

Validate the Deployment

Check that all pods are running:

<code>$ k -n obs-system get pod | grep monitor
alertmanager-monitor-0          2/2     Running   0   3h15m
monitor-kube-state-metrics-...  1/1     Running   0   5h1m
... (other pods) ...</code>

Verify that Prometheus has discovered its targets (screenshot below).

Conclusion

Prometheus Operator represents a shift from manual, passive monitoring to declarative, automatic service discovery; combined with ServiceMonitor it enables “monitoring as soon as a service is deployed.”