kubectl-ai Troubleshooter Skill

Purpose

This skill provides intelligent kubectl-ai prompts for diagnosing and resolving common Kubernetes issues, including pod crashes, image pull errors, resource OOM issues, and service connectivity problems. It suggests fixes via Helm values overrides.

Capabilities

• Generate intelligent kubectl-ai prompts for common Kubernetes issues
• Diagnose Pod crashloop issues with root cause analysis
• Identify and resolve Image pull errors
• Address Resource OOM (Out of Memory) problems
• Troubleshoot Service not reachable issues
• Suggest fixes via Helm values overrides
• Provide remediation commands for identified issues

Implementation Details

Pod CrashLoopBackOff Diagnosis

• Analyze pod logs for crash patterns
• Check container exit codes and reasons
• Examine resource constraints causing crashes
• Identify configuration issues leading to crashes
• Generate appropriate kubectl-ai diagnostic prompts

Image Pull Error Resolution

• Identify registry authentication issues
• Check image name and tag correctness
• Verify pull secrets configuration
• Analyze image pull policy settings
• Generate kubectl-ai prompts for image-related issues

Resource OOM Troubleshooting

• Analyze memory usage patterns
• Check resource limits and requests
• Identify memory leaks in applications
• Suggest appropriate resource adjustments
• Generate kubectl-ai prompts for resource analysis

Service Connectivity Issues

• Check Service and Endpoint configurations
• Verify NetworkPolicies affecting connectivity
• Analyze Ingress configurations if applicable
• Examine DNS resolution issues
• Generate kubectl-ai prompts for networking issues

Helm Values Override Suggestions

• Identify parameters that can be adjusted via Helm values
• Suggest resource limit increases
• Recommend replica count adjustments
• Propose configuration changes via values override
• Generate appropriate Helm upgrade commands

Usage

Common kubectl-ai Prompts for Issues:

Pod CrashLoopBackOff:

kubectl-ai "Analyze pod {pod-name} in namespace {namespace} showing CrashLoopBackOff. Check logs, describe pod, and identify the root cause."

Image Pull Error:

kubectl-ai "Investigate ImagePullBackOff for pod {pod-name}. Check image name, registry access, and pull secrets."

Resource OOM Issues:

kubectl-ai "Examine pod {pod-name} for OutOfMemory issues. Analyze resource usage and suggest appropriate limits."

Service Not Reachable:

kubectl-ai "Troubleshoot why service {service-name} is not reachable. Check endpoints, network policies, and ingress."

Diagnostic Commands:

# Generic troubleshooting
kubectl-ai "Analyze the current state of pods in {namespace} and identify any issues."

# Pod-specific diagnostics
kubectl-ai "Describe pod {pod-name} and explain any abnormal conditions."

# Service connectivity
kubectl-ai "Check connectivity for service {service-name} and suggest fixes."

# Resource analysis
kubectl-ai "Analyze resource usage for deployment {deployment-name} and suggest optimizations."

Helm Values Override Examples:

# Increase memory limits
helm upgrade {release-name} {chart-name} --set backend.resources.limits.memory=512Mi

# Increase CPU limits
helm upgrade {release-name} {chart-name} --set frontend.resources.limits.cpu=500m

# Adjust replica count
helm upgrade {release-name} {chart-name} --set backend.replicaCount=2

# Fix image pull issues
helm upgrade {release-name} {chart-name} --set image.pullSecrets={secret-name}

# Adjust liveness/readiness probes
helm upgrade {release-name} {chart-name} \\
  --set backend.livenessProbe.initialDelaySeconds=60 \\
  --set backend.readinessProbe.periodSeconds=10

Issue-Specific Remediation Patterns:

For CrashLoopBackOff:

1. Check logs:

   kubectl-ai "Show logs from pod {pod-name} --previous"

2. Describe pod:

   kubectl-ai "Describe pod {pod-name} and explain the events"

3. Suggest fix:

   kubectl-ai "Based on the crash reason, suggest Helm values to fix this"

For Image Pull Errors:

1. Check event messages:

   kubectl-ai "Explain the events in pod {pod-name} related to image pulling"

2. Verify registry:

   kubectl-ai "Verify if image {image-name} exists in the registry"

3. Suggest fix:

   kubectl-ai "Recommend Helm values to fix image pull issue"

For OOM Issues:

1. Analyze usage:

   kubectl-ai "Analyze resource usage of pod {pod-name}"

2. Check limits:

   kubectl-ai "Compare current limits with actual usage for pod {pod-name}"

3. Suggest adjustment:

   kubectl-ai "Recommend increased memory limits via Helm values"