Content
# ChaosPilot – AI-Powered Log Analysis Platform
> **Modern, full-stack AI platform for automated log analysis, incident classification, and intelligent remediation.**
---
[🚀 Quick Start](QUICK_START.md) | [📖 Setup Guide](docs/setup-and-deployment/HOW_TO_RUN_AND_DEPLOY_THE_APPLICATION.md) | [🔧 ADK Integration](docs/setup-and-deployment/ADK_INTEGRATION_README.md) | [🧪 Test Integration](test_adk_integration.py)
---
## 🎯 What is ChaosPilot?
ChaosPilot is an intelligent log analysis platform that uses AI agents to automatically analyze error, warning, and critical logs, detect patterns, classify incidents, and recommend fixes.
**We manage the chaos of production errors through intelligent log analysis.**
---
## 🏆 Features
- **🔍 AI-Powered Analysis:** LLMs analyze logs for patterns and anomalies
- **📊 Incident Classification:** Auto-classifies by severity, impact, urgency
- **📋 Smart Response Planning:** Generates actionable response plans
- **🛠️ Automated Fix Recommendations:** Suggests and executes safe fixes
- **⚡ Safe Auto-Fixing:** Automated fixes with rollback capability
- **📢 Smart Alerting:** Intelligent notifications and escalations
- **📈 Real-Time Dashboard:** Live metrics, workflow visualization, and insights
---
## 🏗️ Architecture
| Layer | Tech/Tools |
|------------|-------------------------------------------------|
| Frontend | Angular, TypeScript, TailwindCSS, RxJS |
| Backend | Python, FastAPI, Google ADK, async/await |
| Auth | Supabase (user/session management) |
| Data/AI | Google ADK agents, BigQuery, LLMs |
| DevOps | Docker, GCP, `uv`, `hatch` |
---
## 🚀 Quick Start
### Prerequisites
- [Google Cloud CLI](https://cloud.google.com/sdk/docs/install-sdk)
- [Python 3.9+](https://wiki.python.org/moin/BeginnersGuide/Download)
- [Uv](https://pypi.org/project/uv/) - 🚀 A single tool to replace pip, pip-tools, pipx, poetry, pyenv, twine, virtualenv, and more.
>
- [Node.js 16+](https://nodejs.org/)
- [MCP Toolbox for Databases](https://googleapis.github.io/genai-toolbox/getting-started/local_quickstart/)
### Setup
1. **Clone the repository**
```bash
git clone https://github.com/pmutua/ChaosPilot
cd ChaosPilot
```
2. **Create and activate a Python virtual environment**
- Create virtual environment
```bash
uv venv
```
- Activate virtual environment
```bash
# Linux/macOS:
source .venv/bin/activate
# Windows
.venv\Scripts\activate
```
3. **Install dependencies**
```bash
uv sync
```
4. **Rename `.env.template` to `.env` and add relevant environment variables:**
```env
APP_NAME="agent_manager"
VERSION=0.0.0
MODEL=""
# If you are using Google Gemini use this
GOOGLE_GENAI_USE_VERTEXAI=FALSE
GOOGLE_API_KEY=PASTE_YOUR_ACTUAL_API_KEY_HERE
# If you are using Azure Open AI use this
AZURE_API_KEY="add api key"
AZURE_API_BASE=https://example.openai.azure.com/
AZURE_API_VERSION="2025-05-05-preview"
```
5. **Set up Google Cloud, BigQuery, and MCP Toolbox**
(See [Setup Guide](docs/setup-and-deployment/HOW_TO_RUN_AND_DEPLOY_THE_APPLICATION.md) for full details.)
---
## 🖥️ Running the Application
### Start All Services
```bash
# Start MCP Toolbox
cd mcp-toolbox
toolbox --tools-file="tools.yaml"
# Start ADK API Server (with CORS)
cd ../agent_manager
adk api_server --allow_origins="*"
# or in root directory run:
adk api_server agent_manager --allow_origins="*"
# Start Frontend
cd ../web
npm install
npm start
```
## Debugging with Agent Development Kit UI
Simply run the following command:
```bash
adk web
# And navigate to port 8000. Now you can interact with the Agents and debug.
```
> **NOTE:** Make sure that you select "agent_manager" on "Select Agent" option.
---
## 📚 Documentation
- [Setup & Deployment Guide](docs/setup-and-deployment/HOW_TO_RUN_AND_DEPLOY_THE_APPLICATION.md)
---
## 📊 Usage Examples
### Analyze Error Logs
1. Go to the Log Analysis page
2. Select the "Log Analyzer" agent
3. Provide error logs or describe the issue
4. Get AI-powered analysis with confidence scores
### Get Fix Recommendations
1. Use the "Fix Recommender" agent
2. Share the analyzed logs and issues
3. Get specific fix suggestions with implementation steps
### Automated Incident Response
1. System detects critical issues
2. AI agents generate response plans
3. Safe automated fixes are applied
4. Teams are notified with detailed reports
---
## 🚀 Production Deployment
See [Setup & Deployment Guide](docs/setup-and-deployment/HOW_TO_RUN_AND_DEPLOY_THE_APPLICATION.md) for GCP/Cloud Run instructions.
---
## 🔧 Troubleshooting
- **CORS Errors:** Use `adk api_server app --allow_origins="*"`
- **Service Account Issues:** Run the IAM role assignment scripts
- **Billing Errors:** Ensure GCP billing is enabled
- **Agent Not Found:** Verify you're running from the correct directory
---
## 🤝 Contributing
1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request
---
## 📄 License
MIT License – see the LICENSE file for details.
---
## 🆘 Support
- Create an issue in the repository
- Check the [documentation](docs/setup-and-deployment/)
- Review the [troubleshooting guide](docs/setup-and-deployment/HOW_TO_RUN_AND_DEPLOY_THE_APPLICATION.md#-troubleshooting)
---
**ChaosPilot** – Managing the chaos of production errors through AI-powered log analysis.
