Deploying Swift on Kubernetes: A Basic Practical Guide

Deploying Swift on Kubernetes: A Basic Practical Guide

Swift is not just a language for mobile applications. Thanks to frameworks like Vapor and Kitura, Swift is actively used for building server-side solutions. However, to run such an application in production, reliable orchestration is required. Kubernetes (K8s) is the ideal tool for managing containerized services. In this guide, we will walk through the complete cycle of deploying a Swift application in a Kubernetes cluster: from containerization to creating working manifests.

1. Preparing the Swift Application and Containerization

Before talking about Kubernetes, you need to package your Swift application into a Docker image. Let's look at an example based on the Vapor framework (but the approach is universal for any Swift server).

Create a minimal Vapor application:

// Sources/App/routes.swiftimport Vapor

func routes(_ app: Application) throws {    app.get { req async in        "Hello, Kubernetes!"    }}

Now let's write a Dockerfile. Important: use the official Swift image for building, and a lightweight image (e.g., ubuntu:jammy) for the final container. This will reduce the size and increase security.

# ---- Build Stage ----FROM swift:5.9-jammy AS build

WORKDIR /appCOPY Package.swift Package.resolved ./COPY Sources ./SourcesRUN swift build -c release --static-swift-stdlib

# ---- Run Stage ----FROM ubuntu:jammy

RUN apt-get update && apt-get install -y ca-certificates && rm -rf /var/lib/apt/lists/*WORKDIR /appCOPY --from=build /app/.build/release/Run .

EXPOSE 8080CMD ["./Run"]

Build the image and push it to a registry (e.g., Docker Hub):

docker build -t yourdockerhub/swift-k8s-demo:latest .docker push yourdockerhub/swift-k8s-demo:latest

2. Kubernetes Basics: Pod and Deployment

Kubernetes manages containers through objects. The simplest is a Pod, but for production it's better to use a Deployment, which provides replication and updates.

Create a file deployment.yaml:

apiVersion: apps/v1kind: Deploymentmetadata:  name: swift-demo  labels:    app: swift-demospec:  replicas: 2  selector:    matchLabels:      app: swift-demo  template:    metadata:      labels:        app: swift-demo    spec:      containers:      - name: swift-app        image: yourdockerhub/swift-k8s-demo:latest        ports:        - containerPort: 8080        resources:          requests:            memory: "128Mi"            cpu: "250m"          limits:            memory: "256Mi"            cpu: "500m"        livenessProbe:          httpGet:            path: /            port: 8080          initialDelaySeconds: 5          periodSeconds: 10

Apply the manifest:

kubectl apply -f deployment.yaml

Check the status:

kubectl get pods -l app=swift-demo

3. Service for Accessing the Application

Pods are dynamic (their IPs change upon restart). To allow clients to access the application stably, a Service is needed. Create service.yaml:

apiVersion: v1kind: Servicemetadata:  name: swift-servicespec:  selector:    app: swift-demo  ports:    - protocol: TCP      port: 80      targetPort: 8080  type: ClusterIP

For testing in a local cluster (e.g., Minikube), you can use NodePort or port forwarding:

kubectl port-forward service/swift-service 8080:80

Now the application is available at http://localhost:8080.

4. Scaling and Rolling Updates

One of the main advantages of Kubernetes is easy sc