Kubernetes — Deployment Manifests

Core Principles

• Per-environment manifests: dev-manifest.yaml, staging-manifest.yaml, prod-manifest.yaml
• Token placeholders (#{TOKEN}#) replaced during CI/CD deployment
• Health check probes: /health/ready (readiness), /health/live (liveness)
• RollingUpdate strategy with maxUnavailable: 0, maxSurge: 1
• Environment variables from Secrets and ConfigMaps
• Service Bus topics/subscriptions configured via environment

Key Patterns

Deployment Manifest

apiVersion: apps/v1
kind: Deployment
metadata:
  name: {domain}-command
  namespace: {company}-{env}
spec:
  replicas: #{REPLICAS}#
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 0
      maxSurge: 1
  selector:
    matchLabels:
      app: {domain}-command
  template:
    metadata:
      labels:
        app: {domain}-command
    spec:
      containers:
        - name: {domain}-command
          image: #{ACR_NAME}#.azurecr.io/{domain}-command:#{IMAGE_TAG}#
          ports:
            - containerPort: 8080
            - containerPort: 8081
          env:
            - name: ASPNETCORE_ENVIRONMENT
              value: "#{ENVIRONMENT}#"
            - name: ConnectionStrings__DefaultConnection
              valueFrom:
                secretKeyRef:
                  name: {domain}-secrets
                  key: db-connection-string
            - name: ServiceBus__ConnectionString
              valueFrom:
                secretKeyRef:
                  name: {domain}-secrets
                  key: servicebus-connection
            - name: ExternalServices__QueryServiceUrl
              value: "http://{domain}-query:8081"
          readinessProbe:
            httpGet:
              path: /health/ready
              port: 8080
            initialDelaySeconds: 10
            periodSeconds: 15
          livenessProbe:
            httpGet:
              path: /health/live
              port: 8080
            initialDelaySeconds: 30
            periodSeconds: 30
          resources:
            requests:
              cpu: 100m
              memory: 256Mi
            limits:
              cpu: 500m
              memory: 512Mi

Service

apiVersion: v1
kind: Service
metadata:
  name: {domain}-command
  namespace: {company}-{env}
spec:
  selector:
    app: {domain}-command
  ports:
    - name: http
      port: 8080
      targetPort: 8080
    - name: grpc
      port: 8081
      targetPort: 8081
  type: ClusterIP

Secret (Template)

apiVersion: v1
kind: Secret
metadata:
  name: {domain}-secrets
  namespace: {company}-{env}
type: Opaque
stringData:
  db-connection-string: "#{DB_CONNECTION_STRING}#"
  servicebus-connection: "#{SERVICEBUS_CONNECTION}#"

ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: {domain}-config
  namespace: {company}-{env}
data:
  Serilog__SeqUrl: "http://seq:5341"
  Serilog__WriteToConsole: "true"

Anti-Patterns

Anti-Pattern	Correct Approach
Secrets in ConfigMap	Use Secret for sensitive data
No resource limits	Set requests and limits for stability
Missing health probes	Add readiness and liveness probes
Hardcoded image tags	Use token placeholders for CI/CD
maxUnavailable > 0	Zero downtime: maxUnavailable=0, maxSurge=1

Detect Existing Patterns

# Find K8s manifests
find . -name "*-manifest.yaml" -o -name "*.k8s.yaml" | head -10

# Find token placeholders
grep -r "#{.*}#" --include="*.yaml" .

# Find health check endpoints
grep -r "/health" --include="*.cs" src/

Adding to Existing Project

1. Follow existing manifest naming — {env}-manifest.yaml
2. Match namespace convention — {company}-{env}
3. Use existing secret names for shared resources
4. Add inter-service URLs via environment variables
5. Configure health probes matching the application health endpoints