Kubernetes Manifests

Kubernetes Manifest Authoring

Write and review production-ready Kubernetes manifests following security, resource management, and operational best practices.

Workflow

1. Identify the application type: .NET API, React SPA, worker service, or other
2. Determine the target resources: Deployment, Service, Ingress, HPA, PDB, NetworkPolicy
3. Generate manifests using the patterns below
4. Apply the security context, resource limits, and health probes
5. Validate against the Pre-Deployment Checklist

Core Manifest Structure

Every manifest includes these four required fields:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
  namespace: production
  labels:
    app.kubernetes.io/name: my-app
    app.kubernetes.io/version: "1.0.0"
    app.kubernetes.io/component: api
    app.kubernetes.io/managed-by: kustomize
spec:
  # ...

Use declarative YAML with kubectl apply. Store manifests in version control. Express desired state only.

Security Context

Apply to every pod and container. Non-negotiable for production.

spec:
  securityContext:
    runAsNonRoot: true
    runAsUser: 1000
    runAsGroup: 1000
    fsGroup: 1000
  containers:
  - name: app
    securityContext:
      allowPrivilegeEscalation: false
      readOnlyRootFilesystem: true
      capabilities:
        drop:
          - ALL

For .NET images, use UID 1654 (the app user). For nginx-unprivileged, use UID 101.

Resource Management

Always set memory requests and limits. Set CPU requests; omit CPU limits unless required.

resources:
  requests:
    memory: "256Mi"
    cpu: "100m"
  limits:
    memory: "512Mi"

Set requests = limits for critical services to achieve Guaranteed QoS (lowest eviction priority).

Health Probes

Configure all three probe types. Use dedicated endpoints.

livenessProbe:
  httpGet:
    path: /healthz
    port: 8080
  initialDelaySeconds: 10
  periodSeconds: 15
  timeoutSeconds: 5
  failureThreshold: 3
readinessProbe:
  httpGet:
    path: /ready
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 10
  timeoutSeconds: 5
  failureThreshold: 3
startupProbe:
  httpGet:
    path: /startup
    port: 8080
  failureThreshold: 30
  periodSeconds: 10

Use startupProbe for slow-starting apps instead of large initialDelaySeconds. Keep probes lightweight.

For .NET apps: /health/live, /health/ready, /health/startup. For React/Nginx: /health for both liveness and readiness.

Labels

Apply the app.kubernetes.io/ recommended labels to all resources:

labels:
  app.kubernetes.io/name: myapp
  app.kubernetes.io/instance: myapp-production
  app.kubernetes.io/version: "1.2.3"
  app.kubernetes.io/component: frontend
  app.kubernetes.io/part-of: my-platform
  app.kubernetes.io/managed-by: kustomize

Add custom labels for cross-cutting concerns: environment, team, cost-center, tier.

Container Images

• Never use :latest in production. Pin semantic versions or git SHAs.
• Use digest references (@sha256:...) for maximum reproducibility.
• Use distroless/minimal base images.
• Set imagePullPolicy: IfNotPresent for tagged images.

Graceful Shutdown

Add a preStop hook to handle the delay between endpoint removal and ingress controller configuration updates:

lifecycle:
  preStop:
    exec:
      command: ["sleep", "10"]
terminationGracePeriodSeconds: 45

terminationGracePeriodSeconds must exceed preStop delay + application shutdown timeout.

High Availability

Topology Spread

Distribute pods across zones:

topologySpreadConstraints:
- maxSkew: 1
  topologyKey: topology.kubernetes.io/zone
  whenUnsatisfiable: ScheduleAnyway
  labelSelector:
    matchLabels:
      app: myapp

Pod Disruption Budget

Protect against voluntary disruptions (node drains, upgrades):

apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
  name: myapp-pdb
spec:
  minAvailable: "50%"
  selector:
    matchLabels:
      app: myapp

Use percentages for scaling flexibility. Never set maxUnavailable: 0 or minAvailable: 100%.

HPA (Stateless Apps)

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: myapp-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: myapp
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70
  behavior:
    scaleDown:
      stabilizationWindowSeconds: 300

Services and Ingress

Use ClusterIP for inter-service communication. Use Ingress over multiple LoadBalancer services.

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: myapp-ingress
  annotations:
    nginx.ingress.kubernetes.io/ssl-redirect: "true"
spec:
  ingressClassName: nginx
  tls:
  - hosts:
    - myapp.example.com
    secretName: myapp-tls
  rules:
  - host: myapp.example.com
    http:
      paths:
      - path: /api
        pathType: Prefix
        backend:
          service:
            name: api-service
            port:
              number: 80
      - path: /
        pathType: Prefix
        backend:
          service:
            name: frontend-service
            port:
              number: 80

ConfigMaps and Secrets

Use ConfigMaps for non-sensitive config. Use Secrets for sensitive data. Prefer volume mounts over environment variables. Enable encryption at rest for Secrets.

volumes:
- name: config-volume
  configMap:
    name: app-config
- name: secret-volume
  secret:
    secretName: app-secrets

Use immutable ConfigMaps (immutable: true) for stable configurations. Keep ConfigMaps under 1 MiB.

Deployment Strategy

strategy:
  type: RollingUpdate
  rollingUpdate:
    maxSurge: 25%
    maxUnavailable: 25%

Application-Specific Guidance

For detailed .NET and React/Vite manifest patterns, see:

• references/dotnet-manifests.md — GC configuration, Kestrel ports, structured logging, chiseled images
• references/react-manifests.md — NGINX SPA routing, runtime env vars, non-root NGINX, CSP headers
• references/kustomize.md — Base/overlay structure, generators, patches, components, GitOps integration

Pre-Deployment Checklist

• Resource requests and limits defined (memory required; CPU requests required, limits optional)
• All three health probes configured (liveness, readiness, startup)
• Security context applied: non-root, no privilege escalation, read-only root filesystem
• Labels applied: app.kubernetes.io/name, version, component, managed-by
• Specific image tags pinned (never :latest)
• ConfigMaps for config, Secrets for sensitive data (mounted as volumes)
• PodDisruptionBudget defined for critical workloads
• Topology spread or anti-affinity configured for HA
• preStop hook and terminationGracePeriodSeconds configured for graceful shutdown
• Manifests validated with kubeconform -strict -summary