Connecting To Prometheus

This guide explains how to access the Prometheus server running inside your Kubernetes cluster, including quick diagnostic commands and common access methods (temporary and production-ready). Follow the sequence below to inspect the service, confirm its configuration, and choose an access strategy.

Inspect the Prometheus services

Start by listing services in the cluster to find the Prometheus-related services and their types:

kubectl get service
NAME                                         TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                        AGE
alertmanager-operated                        ClusterIP   None             <none>        9093/TCP,9094/TCP,9094/UDP     11h
kubernetes                                   ClusterIP   10.100.0.1       <none>        443/TCP                        4d5h
prometheus-grafana                           ClusterIP   10.100.235.247   <none>        80/TCP                         11h
prometheus-kube-prometheus-alertmanager      ClusterIP   10.100.53.114    <none>        9093/TCP                       11h
prometheus-kube-prometheus-operator          ClusterIP   10.100.196.32    <none>        443/TCP                        11h
prometheus-kube-prometheus-prometheus        ClusterIP   10.100.54.169    <none>        9090/TCP                       11h
prometheus-kube-state-metrics                ClusterIP   10.100.133.149   <none>        8080/TCP                       11h
prometheus-operated                          ClusterIP   10.100.248.61    <none>        9100/TCP                       11h
prometheus-prometheus-node-exporter          ClusterIP   10.100.249.161   <none>        9100/TCP                       11h

Note the TYPE column — in this example the Prometheus server is exposed via a ClusterIP service (prometheus-kube-prometheus-prometheus), which by default is accessible only from within the cluster.

If you are running commands in a specific namespace other than default, add -n <namespace> to the kubectl commands above. For example: kubectl get service -n monitoring.

Inspect the service YAML

Export the service YAML to confirm its configured ports, selectors, and type:

kubectl get service prometheus-kube-prometheus-prometheus -o yaml > service.yaml

Example excerpt showing the key sections:

apiVersion: v1
kind: Service
metadata:
  annotations:
    meta.helm.sh/release-name: prometheus
    meta.helm.sh/release-namespace: default
  labels:
    app: kube-prometheus-stack-prometheus
    app.kubernetes.io/instance: prometheus
spec:
  clusterIP: 10.100.54.169
  ports:
    - name: http-web
      port: 9090
      protocol: TCP
      targetPort: 9090
  selector:
    app.kubernetes.io/name: prometheus
    prometheus: prometheus-kube-prometheus-prometheus
  type: ClusterIP

The type: ClusterIP confirms the service is internal-only.

How to connect to Prometheus (options)

Choose one of the following access methods depending on your needs. Summary below:

Method	When to use	Pros	Cons	Example command(s)
Port-forward (temporary)	Quick local access for debugging/demos	Fast, no cluster config changes	Not suitable for production; single-user	`kubectl port-forward <pod-name> 9090:9090` or `kubectl port-forward svc/prometheus-kube-prometheus-prometheus 9090:9090`
NodePort	Quick cluster-wide access from node IPs	Simple, exposes on node port	Requires opening node port; not cloud LB-managed	`kubectl edit svc prometheus-kube-prometheus-prometheus` → set `type: NodePort`
LoadBalancer	Production-ready in cloud environments	Cloud LB + external IP	Depends on cloud provider; may incur cost	`kubectl edit svc prometheus-kube-prometheus-prometheus` → set `type: LoadBalancer`
Ingress	Route via domain/host with TLS	Clean host-based routing and certs	Requires Ingress controller and additional config	Create Ingress resource that targets the service

Links and references:

Kubernetes Service types: https://kubernetes.io/docs/concepts/services-networking/service/
kubectl port-forward: https://kubernetes.io/docs/tasks/access-application-cluster/port-forward-access-application-cluster/

Exposing Prometheus directly to the Internet can reveal sensitive metric data. When enabling external access, use authentication, TLS, and network restrictions (Ingress with auth, or firewall rules).

Quick demo — Port-forward to verify the UI

For a short demo or verification you can port-forward the service (or a pod) to your localhost. This does not change the cluster configuration and is temporary.

List pods to find the Prometheus pod name (or use the service name for kubectl port-forward svc/...):

kubectl get pods -l app.kubernetes.io/name=prometheus

Port-forward the Prometheus pod (or service) to localhost 9090:

# Using a pod (replace <prom-pod> with the actual pod name)
kubectl port-forward <prom-pod> 9090:9090

# Or directly from the service
kubectl port-forward svc/prometheus-kube-prometheus-prometheus 9090:9090

Open your browser at: http://localhost:9090

You should see the Prometheus web UI where you can run queries and explore time series data.

The image shows a Prometheus web interface for querying time series data, with options for enabling features like autocomplete and highlighting. It includes fields for expressions and no data has been queried yet.

This local port-forward is ideal for quick checks and demos. For persistent or multi-user access, choose NodePort, LoadBalancer, or an Ingress-based approach and secure it appropriately.

Example: change service type to NodePort (quick patch)

If you decide to expose Prometheus via NodePort briefly, you can patch the service:

kubectl patch svc prometheus-kube-prometheus-prometheus -p '{"spec": {"type": "NodePort"}}'

After patching, run kubectl get svc prometheus-kube-prometheus-prometheus -o wide to see the assigned NODE-PORT. Then you can access Prometheus at http://<node-ip>:<node-port> (ensure firewall / security groups allow the port). For production setups, prefer Ingress with TLS and authentication, or use a cloud LoadBalancer combined with proper access controls.

Introduction

Observability Fundamentals

Prometheus Fundamentals

PromQL

Dashboarding & Visualization

Application Instrumentation

Service Discovery

Push Gateway

Alerting

Monitoring Kubernetes

Conclusion

Scaling & Long Term Storage

Connecting To Prometheus

Inspect the Prometheus services

Inspect the service YAML

How to connect to Prometheus (options)

Quick demo — Port-forward to verify the UI

Example: change service type to NodePort (quick patch)

Watch Video

​Inspect the Prometheus services

​Inspect the service YAML

​How to connect to Prometheus (options)

​Quick demo — Port-forward to verify the UI

​Example: change service type to NodePort (quick patch)

Watch Video

Inspect the Prometheus services

Inspect the service YAML

How to connect to Prometheus (options)

Quick demo — Port-forward to verify the UI

Example: change service type to NodePort (quick patch)