Skip to main content

Deploying the first image

This guide will walk you through the process of rolling out your first application in your cluster – from deployment and release right through to testing.

Prerequisites​

  • A cluster with the status Running
  • A downloaded configuration file and a verified connection; see Connecting to the cluster
  • kubectl installed

First, check that the connection is working:

kubectl get nodes

All nodes should have the status Ready.

1. Create a deployment​

A deployment specifies which image is to run in how many instances. Create the file deployment.yaml:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: web
  labels:
    app: web
spec:
  replicas: 2
  selector:
    matchLabels:
      app: web
  template:
    metadata:
      labels:
        app: web
    spec:
      containers:
        - name: web
          image: nginx:stable
          ports:
            - containerPort: 80
          resources:
            requests:
              cpu: 100m
              memory: 128Mi
            limits:
              cpu: 500m
              memory: 256Mi
          readinessProbe:
            httpGet:
              path: /
              port: 80
            initialDelaySeconds: 5
            periodSeconds: 10

Key information:

  • replicas – Number of instances. Two replicas ensure that the application remains accessible during upgrades.
  • resources.requests – Resources reserved by the scheduler for placement.
  • resources.limits – Upper limit that a container must not exceed.
  • readinessProbe – Ensures that only ready pods receive traffic.

Roll out the deployment:

kubectl apply -f deployment.yaml

2. Check the rollout​

# Status des Rollouts
kubectl rollout status deployment web

# Laufende Pods
kubectl get pods -l app=web

Expected output:

NAME                   READY   STATUS    RESTARTS   AGE
web-6d4cf56db6-8xk2p   1/1     Running   0          30s
web-6d4cf56db6-tz9vn   1/1     Running   0          30s

If a pod gets stuck in Pending or ImagePullBackOff, the following commands will reveal the cause:

kubectl describe pod <pod-name>
kubectl logs <pod-name>

3. Publish the application​

A service makes the pods accessible via a fixed address.

Warnung

Services of type LoadBalancer are currently not available – no Cloud Controller Manager has been set up to assign external addresses. Such a service remains permanently with EXTERNAL-IP: <pending>.

Instead, use NodePort for external access or kubectl port-forward for local testing.

Option A: NodePort for external access​

A service of type NodePort opens the same port on every worker node. The application can then be accessed via the IP address of any node.

Create the file service.yaml:

apiVersion: v1
kind: Service
metadata:
  name: web
spec:
  type: NodePort
  selector:
    app: web
  ports:
    - port: 80
      targetPort: 80
      nodePort: 30080
      protocol: TCP

How to use:

kubectl apply -f service.yaml

The selected port must be within the range 30000–32767. If nodePort is not specified, Kubernetes automatically assigns a free port from this range.

Check the service and the assigned port:

kubectl get service web
NAME   TYPE       CLUSTER-IP     EXTERNAL-IP   PORT(S)        AGE
web    NodePort   10.43.12.180   <none>        80:30080/TCP   10s

You will then need the IP address of a node:

kubectl get nodes -o wide

Access the application via its Node IP and port:

curl http://<node-ip>:30080
Hinweis

The node IP addresses are within the VPC network. The application is therefore only accessible from the internet if the node has a public address and the port is open in the firewall.

As each node opens the same port, if a node fails, its address is no longer available – without an upstream load balancer, you will need to take this into account in your client configuration or in the DNS.

Option B: Port forwarding for local testing​

For a quick test, local port forwarding is sufficient. The service can be of type ClusterIP (the default if type is not specified):

kubectl port-forward service/web 8080:80

The application can then be accessed via http://localhost:8080. The redirection remains in place for as long as the command is running.

4. Scaling the application​

kubectl scale deployment web --replicas=4

To ensure that the replicas are distributed across different nodes, add topologySpreadConstraints – see High Availability.

5. Tidying up​

To remove test resources that are no longer required, use:

kubectl delete -f service.yaml
kubectl delete -f deployment.yaml
Hinweis

A NodePort occupies the selected port on all worker nodes. Delete any services that are no longer required so that the ports are freed up for other applications.

Next steps​

Last updated on