Skip to content

AI Assistant

k13d integrates an intelligent AI assistant that helps you manage Kubernetes clusters with natural language commands and intelligent analysis.

Overview

The AI assistant provides:

  • Natural Language Commands: Ask questions in plain English
  • Context-Aware Analysis: AI has access to YAML, events, and logs
  • Tool Integration: Executes kubectl, bash, and MCP tools
  • Safety Checks: Dangerous commands require approval
  • Beginner Mode: Simple explanations for complex resources

How It Works

┌─────────────────────────────────────────────────────────────────┐
│                    User Question                                  │
│             "Why is my nginx pod failing?"                        │
└─────────────────────────────┬───────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────┐
│                    AI Agent                                       │
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐              │
│  │ Understand  │─►│ Plan Tools  │─►│   Execute   │              │
│  │  Context    │  │   to Use    │  │    Tools    │              │
│  └─────────────┘  └─────────────┘  └─────────────┘              │
└─────────────────────────────┬───────────────────────────────────┘
                              │
              ┌───────────────┼───────────────┐
              ▼               ▼               ▼
        ┌──────────┐   ┌──────────┐   ┌──────────┐
        │ kubectl  │   │   bash   │   │   MCP    │
        │  tools   │   │  tools   │   │  tools   │
        └──────────┘   └──────────┘   └──────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────┐
│                    AI Response                                    │
│  "The nginx pod is failing because the image 'nginx:invalid'    │
│   cannot be pulled. The error indicates 'ImagePullBackOff'.     │
│   Would you like me to fix this by updating the image tag?"     │
└─────────────────────────────────────────────────────────────────┘

Features

Context-Aware Analysis

When you select a resource and ask the AI to analyze it:

  1. YAML Manifest - Full resource definition
  2. Events - Recent Kubernetes events
  3. Logs - Container logs (for Pods)
  4. Metrics - Resource usage data

Built-in Tools

Tool Description Examples
kubectl Execute kubectl commands kubectl get pods, kubectl describe
bash Run shell commands date, grep, file operations
MCP External tool integration GitHub, databases, custom tools

Safety Features

Commands are classified by risk level:

Level Action Examples
Read-only Auto-approved kubectl get, describe
Write Requires approval kubectl apply, create
Dangerous Warning + approval kubectl delete, drain

Usage

TUI Mode

  1. Press Tab to focus on the AI Panel
  2. Type your question
  3. Press Enter to send
  4. Review tool approval requests (Y/N)

Web Mode

  1. Click the AI Assistant panel
  2. Type your question in the input field
  3. Click "Send" or press Enter
  4. Approve/reject tool requests in the popup

Example Conversations

Troubleshooting

You: Why is my nginx deployment not starting?

AI: Let me check your nginx deployment...
[Tool: kubectl get deployment nginx -o yaml]
[Tool: kubectl get events --field-selector involvedObject.name=nginx]

The nginx deployment has 0/3 replicas ready. The issue is:
- Image pull error: "nginx:v999" does not exist
- All pods are in "ImagePullBackOff" state

To fix this, you can update the image to a valid tag:
kubectl set image deployment/nginx nginx=nginx:latest

Resource Explanation

You: Explain this HPA configuration

AI: This HorizontalPodAutoscaler (HPA) is configured to:
- Target: frontend deployment
- Min replicas: 2
- Max replicas: 10
- Scale trigger: CPU utilization > 50%
- Scale-down stabilization: 300 seconds

This means the HPA will add pods when CPU usage exceeds 50%
and remove pods (with a 5-minute delay) when usage drops.

Quick Actions

You: Scale my api deployment to 5 replicas

AI: I'll scale the api deployment to 5 replicas.
[Approval Required: kubectl scale deployment api --replicas=5]

Configuration

LLM Provider Settings

Configure your LLM provider in ~/.config/k13d/config.yaml:

llm:
  provider: openai
  model: gpt-4
  api_key: ${OPENAI_API_KEY}

# Or use Ollama for local LLM
llm:
  provider: ollama
  model: llama3.2
  endpoint: http://localhost:11434

Beginner Mode

Enable beginner mode for simpler explanations:

beginner_mode: true

Best Practices

  1. Be Specific: "Check nginx pod logs for errors" is better than "check pods"
  2. Provide Context: Include namespace and resource names
  3. Review Approvals: Always review tool commands before approving
  4. Use Beginner Mode: Enable for learning or simpler explanations

Next Steps