Just-in

Kagent Tutorial: Build Cloud-Native AI Agents on Kubernetes

Running Ollama LLM on Kubernetes: A Complete Guide

From Docker to Kubernetes: The Orchestration Journey You Didn’t Know You Needed

What is Kubeflow and how can it be used?

Getting Started with Kubeflow: A Complete Beginner’s Guide

Agentic AI Agents AI Cloud-Native AI Kubernetes

What’s New in K8sGPT?

What is K8sGPT?

K8sGPT is an AI-powered command-line tool that scans Kubernetes clusters, diagnoses issues, and provides intelligent recommendations in plain English. Think of it as having an experienced Site Reliability Engineer (SRE) available 24/7 to help troubleshoot your cluster problems.

With over 7,300 GitHub stars and a thriving community, K8sGPT supports multiple AI backends including OpenAI, Azure OpenAI, Google Gemini, Amazon Bedrock, Cohere, and local models through LocalAI or Ollama.

GitHub URL: https://github.com/k8sgpt-ai/k8sgpt

Breaking News: MCP v2 Integration

The latest v0.4.27 release introduces Model Context Protocol (MCP) v2 support github, marking a significant milestone in K8sGPT’s evolution. This integration enables seamless communication between AI assistants and Kubernetes environments through a standardized protocol.

Key MCP Features:

  • Standardized interface for AI-to-Kubernetes communication
  • Natural language queries for cluster operations
  • Streamable HTTP support for real-time interactions
  • Integration with popular AI assistants

What is MCP?

The Model Context Protocol functions as a standardized interface framework that enables AI assistants to communicate with external tools and data sources Medium. For Kubernetes users, this means you can now interact with your clusters using conversational AI without memorizing complex kubectl commands.

Claude Desktop Integration: Game-Changing AI-Powered Cluster Management

One of the most exciting developments is K8sGPT’s native integration with Claude Desktop, requiring version 0.4.14 or later github. This feature transforms how teams interact with Kubernetes clusters.

How to Enable Claude Desktop Integration

# Start K8sGPT MCP server
k8sgpt serve --mcp --mcp-http

# Custom port configuration
k8sgpt serve --mcp --mcp-http --mcp-port 8089

# Full serve mode with metrics
k8sgpt serve --mcp --mcp-http --port 8080 --metrics-port 8081 --mcp-port 8089

Configure Claude Desktop

Add this configuration to your Claude Desktop settings:

{
  "mcpServers": {
    "k8sgpt": {
      "command": "k8sgpt",
      "args": [
        "serve",
        "--mcp"
      ]
    }
  }
}

What You Can Do:

  • “Analyze my Kubernetes cluster for issues”
  • “What pods are failing in the production namespace?”
  • “Show me resource utilization across all nodes”
  • “Explain why my deployment is not scaling”

New Operator Lifecycle Management (OLM) Analyzers

Version 0.4.24 introduced comprehensive OLM support with new analyzers for ClusterServiceVersion, Subscription, InstallPlan, OperatorGroup, and CatalogSource github. This is crucial for organizations using Operator Framework to manage Kubernetes operators.

Additionally, version 0.4.23 added ClusterCatalog and ClusterExtension analyzers github, providing complete visibility into operator deployments.

OLM Analyzers Include:

  • ClusterServiceVersion (CSV): Validates operator installations and metadata
  • Subscription: Checks operator update channels and versions
  • InstallPlan: Monitors operator installation and upgrade plans
  • OperatorGroup: Ensures proper namespace targeting
  • CatalogSource: Validates operator catalog availability
  • ClusterCatalog: Analyzes cluster-wide operator catalogs
  • ClusterExtension: Reviews operator extensions and add-ons

Enhanced Cloud Provider Support

K8sGPT continues expanding its multi-cloud capabilities with significant enhancements:

Amazon Bedrock Improvements

Version 0.4.22 added support for APAC region Claude models and inference profiles github, allowing organizations in Asia-Pacific to leverage Claude models with lower latency.

Inference Profile Configuration:

# System Inference Profile
k8sgpt auth add --backend amazonbedrock \
  --providerRegion us-east-1 \
  --model arn:aws:bedrock:us-east-1:123456789012:inference-profile/my-inference-profile

# Application Inference Profile
k8sgpt auth add --backend amazonbedrock \
  --providerRegion us-east-1 \
  --model arn:aws:bedrock:us-east-1:123456789012:application-inference-profile/2uzp4s0w39t6

Claude 4 Support

Version 0.4.21 introduced support for Claude 4 models and updated model name listings github, ensuring users have access to the latest and most capable AI models.

Oracle Cloud Infrastructure (OCI)

Version 0.4.23 brought OCI GenAI chat models support github, expanding K8sGPT’s reach to Oracle Cloud users.

Advanced Cache Management

Version 0.4.20 introduced the cache purge feature github, giving administrators fine-grained control over cached analysis results.

Cache Management Commands:

# List cached items
k8sgpt cache list

# Purge specific cached object
k8sgpt cache purge $OBJECT_NAME

# Add remote cache (AWS S3)
k8sgpt cache add s3 --region us-east-1 --bucket my-k8sgpt-cache

# Add remote cache (Azure Blob Storage)
k8sgpt cache add azure \
  --storageacc mystorageaccount \
  --container k8sgpt-cache

# Add remote cache (Google Cloud Storage)
k8sgpt cache add gcs \
  --region us-central1 \
  --bucket my-k8sgpt-cache \
  --projectid my-project

# Remove remote cache
k8sgpt cache remove

Custom Analyzer Framework

K8sGPT now supports custom analyzers, allowing organizations to extend functionality for proprietary Kubernetes resources or custom operators.

Custom Analyzer Configuration:

custom_analyzers:
  - name: host-analyzer
    connection:
      url: localhost
      port: 8080

Commands:

# List custom analyzers
k8sgpt custom-analyzer list

# Add custom analyzer
k8sgpt custom-analyzer add --name my-analyzer --port 8085

# Remove custom analyzers
k8sgpt custom-analyzer remove --names "analyzer1,analyzer2"

# Run analysis with custom analyzers
k8sgpt analyze --custom-analysis

Latest Model Support

K8sGPT now supports an impressive array of AI models and providers:

Supported Backends:

  • OpenAI: GPT-4, GPT-4 Turbo, GPT-3.5 Turbo
  • Anthropic: Claude 4 (Opus, Sonnet), Claude 3.5
  • Azure OpenAI: Enterprise-grade GPT models
  • Google: Gemini Pro, Gemini Ultra
  • Amazon Bedrock: Claude, Titan, Llama models
  • Cohere: Command, Command-R
  • Oracle Cloud: OCI GenAI models
  • Local Models: Ollama, LocalAI
  • IBM watsonx.ai: Enterprise AI models

Built-in Analyzers

K8sGPT includes 25+ built-in analyzers covering:

Core Resources (Enabled by Default):

  • Pod, Service, Deployment, StatefulSet
  • ReplicaSet, Job, CronJob
  • PersistentVolumeClaim, Node
  • Ingress, ConfigMap
  • MutatingWebhook, ValidatingWebhook

Optional Resources:

  • HorizontalPodAutoscaler (HPA)
  • PodDisruptionBudget (PDB)
  • NetworkPolicy
  • Gateway API (GatewayClass, Gateway, HTTPRoute)
  • Log Analysis, Storage Analysis, Security Analysis
  • OLM Resources (CSV, Subscription, InstallPlan, etc.)

How to Get Started

Installation

# macOS/Linux via Homebrew
brew install k8sgpt

# Or from tap
brew tap k8sgpt-ai/k8sgpt
brew install k8sgpt

# Ubuntu/Debian (64-bit)
curl -LO https://github.com/k8sgpt-ai/k8sgpt/releases/download/v0.4.27/k8sgpt_amd64.deb
sudo dpkg -i k8sgpt_amd64.deb

# RHEL/CentOS/Fedora (64-bit)
sudo rpm -ivh https://github.com/k8sgpt-ai/k8sgpt/releases/download/v0.4.27/k8sgpt_amd64.rpm

# Alpine (64-bit)
wget https://github.com/k8sgpt-ai/k8sgpt/releases/download/v0.4.27/k8sgpt_amd64.apk
apk add --allow-untrusted k8sgpt_amd64.apk

Quick Start

# Generate API key (opens browser)
k8sgpt generate

# Configure authentication
k8sgpt auth add --backend openai --password $OPENAI_API_KEY

# Run analysis
k8sgpt analyze

# Get detailed explanations
k8sgpt analyze --explain

# Include Kubernetes documentation
k8sgpt analyze --explain --with-doc

# Filter by resource type
k8sgpt analyze --filter=Pod,Service --explain

# Filter by namespace
k8sgpt analyze --filter=Pod --namespace=production --explain

# Output as JSON
k8sgpt analyze --explain --output=json

# Anonymize sensitive data
k8sgpt analyze --explain --anonymize

Why K8sGPT Matters in 2025

As Kubernetes clusters grow in complexity, troubleshooting becomes increasingly challenging. K8sGPT addresses this by:

  1. Democratizing Kubernetes Expertise: Making cluster troubleshooting accessible to developers without deep K8s knowledge
  2. Reducing MTTR: Faster issue identification and resolution with AI-powered insights
  3. Continuous Monitoring: Operator mode enables 24/7 cluster health monitoring
  4. Multi-Cloud Support: Works across all major cloud providers and on-premises deployments
  5. Natural Language Interface: Interact with clusters using conversational AI through MCP

Best Practices

  1. Use Anonymization in Production: Always enable --anonymize when sharing analysis results
  2. Configure Remote Caching: Use S3, Azure Blob, or GCS for distributed teams
  3. Enable Selective Analyzers: Focus on relevant resource types to reduce noise
  4. Integrate with CI/CD: Add K8sGPT checks to deployment pipelines
  5. Leverage Custom Analyzers: Extend functionality for custom resources
  6. Enable MCP for AI Assistants: Connect with Claude Desktop or other MCP clients

Community and Resources

Conclusion

K8sGPT’s recent releases represent a significant leap forward in AI-assisted Kubernetes operations. The MCP v2 integration, Claude Desktop support, expanded OLM analyzers, and enhanced cloud provider capabilities make it an indispensable tool for modern DevOps teams.

Whether you’re managing a single cluster or orchestrating multi-cloud Kubernetes deployments, K8sGPT’s latest features provide the intelligence and automation needed to maintain healthy, efficient clusters.

Ready to give your Kubernetes superpowers? Install K8sGPT today and experience the future of cluster management.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *