dockai 0.5.1

DockAI – AI-powered Docker Log Analysis Tool (CLI + Cloud)

🐳 DockAI – AI-powered Docker Log Analysis Tool (CLI + Cloud)

DockAI is an intelligent CLI tool that analyzes Docker container logs using Large Language Models (LLMs). It helps developers, DevOps engineers, and system administrators quickly identify issues, summarize logs, and provide actionable insights.

---

🚀 Features

• AI Log Analysis Understands and summarizes logs using LLMs, identifying possible root causes and suggesting solutions.

• Performance Monitoring (CPU & Memory) Measure container performance in real-time or over a time window using --perf or --instant-perf.

• Local & Cloud AI Modes (Ollama + OpenAI) Analyze with a local model (e.g., llama3) or cloud-based OpenAI API:

  dockai analyze my-container --mode local
  dockai analyze my-container --mode cloud
  

• Live Container Status Even when no logs are generated, DockAI provides a live summary including container status, restart count, and health.

• Simple CLI Usage

  dockai analyze <container-name> --since 15m --tail 3000
  

---

📊 Example Output

🤖 AI Analysis:
**Summary:** Database connection failed.
**Root Cause:** TCP/IP connection refused.
**Solution:** 
- Restart the database service inside the Docker container.
- Check port accessibility and network configuration.

⚙️ Performance
- CPU p95: 0.3% | max: 1.1%
- Mem p95: 12.7%

---

🧠 Supported Models

• Local: Ollama (e.g., llama3, mistral, gemma)
• Cloud: OpenAI GPT-4, GPT-4o-mini

🔧 Default Model

By default, DockAI uses:

DOCKAI_OLLAMA_MODEL = "qwen2.5:7b-instruct"

This model offers excellent multilingual support (including Turkish 🇹🇷) and strong technical reasoning for analyzing Docker logs.

To override the model, set an environment variable:

export DOCKAI_OLLAMA_MODEL="aya:23b"

---

⚙️ Installation

pip install dockai

🧩 Ollama Installation (All Platforms)

macOS

brew install ollama
ollama pull qwen2.5:7b-instruct

Linux

curl -fsSL https://ollama.com/install.sh | sh
ollama pull qwen2.5:7b-instruct

Windows (PowerShell)

winget install Ollama.Ollama
ollama pull qwen2.5:7b-instruct

💡 Tip: DockAI automatically uses the model defined in the environment variable DOCKAI_OLLAMA_MODEL (default: qwen2.5:7b-instruct).

---

🧩 Developer Commands

make build       # build the package
make publish     # publish to PyPI
make testpublish # publish to TestPyPI

---

🧾 License

Apache License 2.0 Copyright (c) 2025 Ahmet Atakan

---

🧩 Plugin Architecture

DockAI supports a modular plugin system that allows developers to extend functionality without modifying the core codebase.
Each plugin can react to lifecycle hooks such as on_start, on_finish, or on_error.

🔌 How Plugins Work

• Plugins are loaded automatically from:
  • dockai/plugins/ (built-in plugins)
  • ~/.dockai/plugins/ (user-installed plugins)
• Each plugin defines a plugin.json file that describes:

  {
    "enabled": true,
    "name": "telemetry",
    "version": "0.2.0",
    "config": {
      "sqlite_path": "~/.dockai/usage.db"
    }
  }
  

✨ Example Plugin Hooks

def on_start(self, ctx):
    print("[plugin] analysis started")

def on_finish(self, ctx):
    print("[plugin] analysis completed")

---

📈 Telemetry & Usage Tracking

DockAI includes a built-in Telemetry Plugin for tracking usage and performance statistics.
This plugin helps monitor how DockAI is used, improving future versions and providing analytics for paid plans.

📊 Data Model

• usage table — stores each analysis run (time, container, mode, latency, etc.)
• findings table — stores detected errors/warnings and AI insights per run

🔒 Privacy

All telemetry data is stored locally in SQLite (~/.dockai/usage.db) and never sent externally.
Users can disable or extend telemetry via plugin configuration.

---

💰 Licensing & Monetization Roadmap

DockAI is open source (Apache 2.0) but designed to support optional monetization:

• Free Plan: limited analysis history and findings
• Pro Plan (License Key): unlocks unlimited telemetry, detailed analytics, and advanced plugins
• Plugin Marketplace (future): third-party verified plugins with SHA-based signature validation

A JWT-based license verification system is planned to allow easy activation via:

dockai license activate --key <YOUR_LICENSE_KEY>

---

🗺️ Future Roadmap

• 🧠 Enhanced AI reasoning & multi-model ensemble
• 📊 Graphical performance reports (PDF or Web)
• 🔐 Secure license key & API-based billing
• 🧩 Plugin Store with auto-update mechanism
• 🌍 Cloud telemetry dashboard

---