Kubernetes Agent

This project implements a Flask-based Kubernetes assistant that interacts with a Kubernetes cluster and responds to user queries. It uses OpenAI's GPT model with tools to provide concise, JSON-formatted responses to Kubernetes-related questions.

Table of Contents

• Features
• Installation
• Prerequisites
• Steps to Installation
• Endpoints
• Checklist
• License

Features

Kubernetes-Agent is an AI powered assistant that will analyze your kubernetes cluster and answer all of the specifics regarding it.

The Agent:

• Handles user queries and provides accurate answers in JSON format
• Retrieves Kubernetes cluster information which includes nodes, pods, deployments and services
• Utilizes OpenAI's Function Callinf for intuitive query handling

By using the power of function calling the LLM can and will determine when to fetch you information about your deployed cluster. All you need to locate is your kubeconfig file its fair game!

Installation

Prerequisites

• Python 3.10 and above.
• Kubernetes cluster is properly configured and running (ensure the location of config file)
• Retrieve OpenAI Key
• Flask is installed

Steps to Installation

1. Clone the repository

git clone <repository-url>
cd <repository-folder>

2. Create a vitual environment and install dependencies

```bash
conda create --name <env_name>
conda activate <env_name>
pip install -r requirements.txt
```

3. Set OpenAI API key

```bash
API_KEY = your_api_key
```

4. Configure the Kubernetes cluster (minikube and kubectl)

• Start minikube

  minikube start

• Make sure kubectl is working

  kubectl get nodes

Running the Application

Start the Backend

Run the Flask application

python main.py

• Note that the application will run on http://0.0.0.0:8000

Endpoints

1. /query

• Methods : POST
• Description: Sends a query to the Agent and receives a response

Example

curl -X POST http://localhost:8000/query \
-H "Content-Type: application/json" \
-d '{"query":"Which pod is spawned by my-deployment?"}'

• Response

{
  "response": {
    "query": "Which pod is spawned by my-deployment?",
    "answer": "my-pod"
  }
}

2. /start_chat

• Methods : POST
• Description: Starts a chat with the Agent, ensuring a unique identifier session_id. For each user currently the session_id is hard coded to 1.

Example

curl -X POST http://localhost:8000/start_chat '

• Response

{
  "response": {
    "message": "You are starting a conversation with our AI Agent!", 
    "session_id": hkljsdj23sd12
    }
}

3. /get_kube_api

• Methods : POST
• Description: Fetches detailed information about the Kubernetes cluster. (This is mainly used internally by the Agent)

Example

curl -X GET http://localhost:8000/get_kube_api'

• Response

{
  "cluster_info": {
    "query": "Get Kubernetes Cluster Info",
    "answer": "{...}"  // JSON representation of cluster details
  }
}

Check before You Start!

• Start Backend: Ensure the backend runs on http://0.0.0.0:8000.
• Endpoints Tested: Confirm /start_chat, /query, and /get_kube_api work as expected.
• Log Debugging: Logs are written to agent.log.
• Concise Responses: Ensure responses follow the format:

{
  "query": "Which pod is spawned by my-deployment?",
  "answer": "my-pod"
}

Have Fun!

License

This project is licensed under the MIT License