Kubernetes Objects¶

Kubernetes objects are persistent entities that represent the state of your cluster.

Understanding Kubernetes Objects¶

What are Objects?¶

Objects are records of intent - once created, Kubernetes works to ensure the object exists.

Key Characteristics: - Persistent entities - Represent desired state - Managed via API - Stored in etcd

Object Spec and Status¶

Every Kubernetes object has two nested fields:

apiVersion: v1
kind: Pod
metadata:
  name: my-pod
spec:              # Desired state
  containers:
  - name: nginx
    image: nginx
status:            # Current state (managed by K8s)
  phase: Running
  conditions: []

Required Fields¶

All objects require these fields:

apiVersion: v1           # API version
kind: Pod                # Object type
metadata:                # Identifying data
  name: my-object
  namespace: default
spec:                    # Desired state
  # Object-specific fields

Object Categories¶

1. Workload Objects¶

Manage application workloads:

Pod: Smallest deployable unit
Deployment: Declarative updates for Pods
StatefulSet: Stateful applications
DaemonSet: One pod per node
Job: Run-to-completion tasks
CronJob: Scheduled jobs

2. Service Objects¶

Expose applications:

Service: Stable network endpoint
Ingress: HTTP/HTTPS routing
EndpointSlice: Network endpoints

3. Storage Objects¶

Manage persistent data:

PersistentVolume (PV): Storage resource
PersistentVolumeClaim (PVC): Storage request
StorageClass: Dynamic provisioning
VolumeSnapshot: Volume backup

4. Configuration Objects¶

Manage configuration:

ConfigMap: Configuration data
Secret: Sensitive data

5. Security Objects¶

Control access:

ServiceAccount: Pod identity
Role: Namespace permissions
ClusterRole: Cluster-wide permissions
RoleBinding: Bind role to subjects
ClusterRoleBinding: Cluster-wide binding

6. Cluster Objects¶

Cluster-level resources:

Namespace: Virtual clusters
Node: Worker machine
PersistentVolume: Cluster storage
ClusterRole: Cluster permissions

Object Management¶

Imperative Commands¶

Quick operations without YAML:

# Create deployment
kubectl create deployment nginx --image=nginx

# Expose service
kubectl expose deployment nginx --port=80

# Scale deployment
kubectl scale deployment nginx --replicas=3

# Delete resource
kubectl delete deployment nginx

Imperative Object Configuration¶

Using YAML files:

# Create from file
kubectl create -f deployment.yaml

# Update from file
kubectl replace -f deployment.yaml

# Delete from file
kubectl delete -f deployment.yaml

Declarative Object Configuration¶

Recommended for production:

# Apply configuration
kubectl apply -f deployment.yaml

# Apply directory
kubectl apply -f ./configs/

# Apply with recursive
kubectl apply -f ./configs/ -R

# View last applied configuration
kubectl get deployment nginx -o yaml

Object Metadata¶

Labels¶

Key-value pairs for organizing objects:

metadata:
  labels:
    app: nginx
    environment: production
    tier: frontend
    version: "1.0"

Label Selectors:

# Equality-based
kubectl get pods -l app=nginx
kubectl get pods -l environment!=production

# Set-based
kubectl get pods -l 'environment in (production, staging)'
kubectl get pods -l 'tier notin (frontend, backend)'

# Multiple selectors
kubectl get pods -l app=nginx,environment=production

Annotations¶

Non-identifying metadata:

metadata:
  annotations:
    description: "Production nginx deployment"
    contact: "devops@example.com"
    buildVersion: "1.2.3"
    kubernetes.io/change-cause: "Update to version 1.2.3"

Namespaces¶

Virtual clusters for resource isolation:

# List namespaces
kubectl get namespaces

# Create namespace
kubectl create namespace dev

# Set default namespace
kubectl config set-context --current --namespace=dev

# Get resources in namespace
kubectl get pods -n dev

# Get resources in all namespaces
kubectl get pods --all-namespaces

Object Names and UIDs¶

Names¶

User-provided identifiers:

metadata:
  name: my-pod        # Must be unique within namespace
  namespace: default

Naming Constraints: - Max 253 characters - Lowercase alphanumeric, -, . - Start and end with alphanumeric

UIDs¶

System-generated unique identifiers:

# View UID
kubectl get pod my-pod -o jsonpath='{.metadata.uid}'

Owner References¶

Establish parent-child relationships:

metadata:
  ownerReferences:
  - apiVersion: apps/v1
    kind: ReplicaSet
    name: nginx-deployment-7d64c5d5c9
    uid: 12345678-1234-1234-1234-123456789012
    controller: true
    blockOwnerDeletion: true

Garbage Collection: - Foreground: Delete dependents first - Background: Delete owner immediately - Orphan: Leave dependents

# Delete with cascade
kubectl delete deployment nginx --cascade=foreground

# Delete without cascade (orphan)
kubectl delete deployment nginx --cascade=orphan

Finalizers¶

Prevent deletion until conditions are met:

metadata:
  finalizers:
  - kubernetes.io/pvc-protection
  - example.com/custom-finalizer

Field Selectors¶

Query objects by field values:

# By status
kubectl get pods --field-selector status.phase=Running

# By metadata
kubectl get pods --field-selector metadata.namespace=default

# Multiple fields
kubectl get pods --field-selector status.phase=Running,spec.restartPolicy=Always

Common Object Patterns¶

Basic Pod¶

apiVersion: v1
kind: Pod
metadata:
  name: nginx-pod
  labels:
    app: nginx
spec:
  containers:
  - name: nginx
    image: nginx:1.21
    ports:
    - containerPort: 80

Deployment¶

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.21
        ports:
        - containerPort: 80

Service¶

apiVersion: v1
kind: Service
metadata:
  name: nginx-service
spec:
  selector:
    app: nginx
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80
  type: LoadBalancer

ConfigMap¶

apiVersion: v1
kind: ConfigMap
metadata:
  name: app-config
data:
  database_url: "postgres://db:5432"
  log_level: "info"

Secret¶

apiVersion: v1
kind: Secret
metadata:
  name: app-secret
type: Opaque
data:
  password: cGFzc3dvcmQxMjM=  # base64 encoded

Object Lifecycle¶

┌──────────┐
│ Created  │
└────┬─────┘
     │
     ▼
┌──────────┐
│ Pending  │
└────┬─────┘
     │
     ▼
┌──────────┐
│ Running  │
└────┬─────┘
     │
     ▼
┌──────────┐
│Succeeded │
│  Failed  │
│ Unknown  │
└────┬─────┘
     │
     ▼
┌──────────┐
│ Deleted  │
└──────────┘

Viewing Objects¶

# Get resources
kubectl get pods
kubectl get deployments
kubectl get services

# Describe resource
kubectl describe pod nginx-pod

# Get YAML
kubectl get pod nginx-pod -o yaml

# Get JSON
kubectl get pod nginx-pod -o json

# Custom columns
kubectl get pods -o custom-columns=NAME:.metadata.name,STATUS:.status.phase

# JSONPath
kubectl get pods -o jsonpath='{.items[*].metadata.name}'

# Watch resources
kubectl get pods --watch

Editing Objects¶

# Edit live object
kubectl edit deployment nginx

# Patch object
kubectl patch deployment nginx -p '{"spec":{"replicas":5}}'

# Set image
kubectl set image deployment/nginx nginx=nginx:1.22

# Annotate
kubectl annotate pod nginx description="Production pod"

# Label
kubectl label pod nginx environment=production

Best Practices¶

Use Declarative Configuration
Store YAML in version control
Use kubectl apply
Enable GitOps workflows
Organize with Labels
Consistent labeling scheme
Use for selection and grouping
Include app, environment, version
Use Namespaces
Separate environments
Resource quotas per namespace
RBAC per namespace
Document with Annotations
Add descriptions
Track changes
Store metadata
Resource Management
Set resource requests/limits
Use resource quotas
Monitor usage