CKAD — Managing Deployments

1 minute read

Editing Deployments

Several ways to modify a Deployment:

# Edit interactively (opens in $EDITOR)
kubectl edit deployment web

# Patch a specific field
kubectl patch deployment web -p '{"spec":{"replicas":5}}'

# Set image
kubectl set image deployment/web nginx=nginx:1.27

# Set resources
kubectl set resources deployment/web -c nginx --requests=cpu=200m,memory=256Mi

# Apply updated YAML
kubectl apply -f web-deployment.yaml

Which Method to Use

Method Best For
kubectl edit Quick interactive changes
kubectl patch Changing a single field programmatically
kubectl set image Updating container images (most common)
kubectl set resources Changing CPU/memory requests and limits
kubectl apply -f Applying a complete updated YAML file

Deployment Annotations and Labels

Add metadata to track changes:

# Annotate a Deployment
kubectl annotate deployment web kubernetes.io/change-cause="Updated nginx to 1.27"

# This shows up in rollout history
kubectl rollout history deployment/web
# REVISION  CHANGE-CAUSE
# 1         <none>
# 2         Updated nginx to 1.27

# Add labels
kubectl label deployment web team=backend env=production

Exam tip: The --record flag (deprecated but still works) automatically records the command as the change cause. You can also use kubectl annotate to set kubernetes.io/change-cause manually.

Monitoring Deployments

Quick Status

# Quick status
kubectl get deployment web

# Detailed with conditions
kubectl describe deployment web

# Watch Pods being updated in real time
kubectl get pods -l app=web -w

Checking ReplicaSets

Each Deployment update creates a new ReplicaSet. The old ones are kept (scaled to 0) for rollbacks:

# Check ReplicaSets (old and new)
kubectl get rs -l app=web
# Expected:
# NAME             DESIRED   CURRENT   READY   AGE
# web-5d9f8b6c4    3         3         3       5m   ← current
# web-7b4f9d3a2    0         0         0       30m  ← previous

Events

# Check events for issues
kubectl get events --sort-by=.lastTimestamp | grep web

# Or use describe for resource-specific events
kubectl describe deployment web | tail -20

Deployment Conditions

kubectl get deployment web -o jsonpath='{.status.conditions[*].type}'
# Common conditions:
# Available    — Minimum replicas are ready
# Progressing  — Rollout is in progress
# ReplicaFailure — Cannot create Pods (e.g., resource limits)

Monitoring Quick Reference

Task Command
Deployment status kubectl get deployment <name>
Detailed info kubectl describe deployment <name>
Watch Pods kubectl get pods -l app=<name> -w
ReplicaSets kubectl get rs -l app=<name>
Current image kubectl describe deployment <name> \| grep Image
Events kubectl get events --sort-by=.lastTimestamp
JSON output kubectl get deployment <name> -o json

Next: Learn about advanced deployment patterns with Blue-Green & Canary →

Updated:

You may also enjoy