K8s Ai

An intelligent system that combines AI with Kubernetes management, enabling cluster diagnosis, resource monitoring, and log analysis through natural language interaction to simplify K8s operations and maintenance work.

Developer tools Monitoring #Intelligent Operations and Maintenance #K8s Management #AI Analysis #Natural Interaction .kotlin

rating : 2.5 points

downloads : 7.1K

update time : 2025-04-28

Open Site

What is the MCP Server?

The MCP Server is the core component of the Kubernetes AI Management System that provides intelligent cluster management through natural language processing. It acts as a bridge between users and their Kubernetes clusters, understanding plain English queries about cluster status, resources, and operations.

How to use the MCP Server?

The MCP Server can be integrated with AI interfaces like Claude Desktop or used directly via REST API. Users simply ask questions in natural language about their Kubernetes cluster, and the system provides detailed responses with cluster information and recommendations.

Use Cases

Ideal for DevOps teams, SREs, and platform engineers who need to monitor and troubleshoot Kubernetes clusters without memorizing complex kubectl commands. Also valuable for developers who need occasional cluster access but aren't Kubernetes experts.

Key Features

Cluster Health Monitoring

Get real-time insights into your cluster's health status, node conditions, and resource utilization

Natural Language Interface

Interact with your cluster using plain English questions instead of memorizing kubectl commands

Resource Analysis

Identify resource hogs, unbalanced workloads, and optimization opportunities in your cluster

Smart Log Analysis

AI-powered log parsing that highlights important patterns and anomalies in pod logs

Helm Release Management

Manage Helm releases, check versions, and perform upgrades/rollbacks through simple commands

Advantages

No need to learn complex kubectl commands - use natural language

Comprehensive cluster visibility through simple queries

AI-powered diagnostics identify issues you might miss

Saves time on routine cluster monitoring tasks

Accessible to non-Kubernetes experts

Limitations

Requires properly configured kubeconfig access

Some advanced operations still require direct kubectl usage

AI interpretation may occasionally need clarification

Initial setup requires Java/Kubernetes knowledge

Getting Started

Prerequisites

Ensure you have Java 17+, Maven 3.8+, and a configured Kubernetes cluster with kubeconfig

Build the Project

Clone the repository and build all modules using Maven

Run the MCP Server

Start the MCP server which will be ready to process your queries

Integration

Connect the MCP server to your preferred interface (Claude Desktop or REST client)

Example Scenarios

Cluster Health Check

Quickly assess the overall health of your Kubernetes cluster

Troubleshooting Failing Pods

Identify and diagnose pods that are crashing or in error state

Resource Optimization

Find resource hogs and optimization opportunities

Frequently Asked Questions

What Kubernetes versions are supported?

Can I use this with managed Kubernetes services like EKS or AKS?

Is there a web interface or only CLI?

How secure is the natural language interface?

Additional Resources

Official Kubernetes Documentation

Comprehensive Kubernetes reference

GitHub Repository

Source code and issue tracking

Spring Boot Documentation

Underlying framework documentation

🚀 Kubernetes AI Management System

Kubernetes management with AI (MCP + Agent)

    ⎈ K8s AI Management
    ├── 🤖 MCP Server
    ├── 🔍 Kubernetes Tools
    └── 🚀 Agent Mode with Rest API

🚀 Quick Start

This project combines artificial intelligence with Kubernetes management. Users can perform real - time diagnosis, resource monitoring, and intelligent log analysis, simplifying Kubernetes management and offering a modern alternative.

💡 Simply ask questions naturally - no need to memorize commands!

✨ Features

This AI - driven system understands natural - language queries about Kubernetes clusters. Here are some of the functions that the system can provide through natural - language queries:

🏥 Cluster Health and Diagnosis

"How is the status of my cluster?"
"Show all pods in the default namespace."
"Are there any failed pods in the default namespace?"
"Is my application running properly?"
"Do I need to optimize the resource usage of my Kubernetes cluster?"

Detailed Functional Modules

🛠️ Cluster Status Monitoring

View the real - time health status of the cluster, including the status of nodes, pods, and services.

🚨 Anomaly Detection

Automatically identify and report potential issues, such as pod failures or resource shortages.

🔍 Resource Usage Analysis

Provide detailed resource usage reports to help optimize the performance of the Kubernetes cluster.

📈 Performance Monitoring

Track key metrics, such as CPU, memory, and disk usage, and generate trend charts.

🛠️ Cluster Status Monitoring

View the real - time operating status of all nodes.
Monitor the health status of pods.
Check if services are running properly.

🚨 Anomaly Detection

Automatically identify pod failures.
Report resource shortage situations.
Provide repair suggestions.

🔍 Resource Usage Analysis

Analyze CPU and memory usage.
Generate resource usage reports.
Help optimize cluster configuration.

📈 Performance Monitoring

Monitor key metrics.
Generate trend charts.
Provide performance optimization suggestions.

📦 Installation

Prerequisites

JDK 8 or higher
Maven 3.5 or higher
Kubernetes environment

Build the Project

# Download the project source code
git clone https://github.com/your-repository.git
cd kubernetes-ai-system

# Build with Maven
mvn clean package

Test the Project

3.1 Integration of MCP Server and Claude Desktop

Refer to mcp-server/README.md for instructions on how to integrate with Claude Desktop.

3.2 Agent Mode with Rest API

Refer to agent/README.md for instructions on how to run the agent.

📚 Documentation

Project Structure

The project is divided into the following modules:

agent: Agent mode with Rest API, used to analyze the cluster through natural language.
mcp-server: MCP server, used in combination with tools (such as Claude Desktop) to provide a complete experience.
tools: Kubernetes tools for cluster analysis/management (used jointly by the agent and mcp - server).