dlia

Project Overview

DLIA (Docker Log Intelligence Agent) is an AI-powered Docker log monitoring agent that uses Large Language Models to intelligently analyze container logs, detect anomalies, and provide contextual insights over time.

Key Features

• Semantic log analysis: LLMs understand log context, not just keywords
• Historical context: Tracks trends over time to detect gradual degradation
• Natural language filtering: Ignore routine errors in plain English
• Self-cleaning knowledge base: Auto-forgets old issues with configurable retention
• Privacy-first: Automatic anonymization of IPs, secrets, and sensitive data
• Flexible LLM backend: OpenAI, OpenRouter, Ollama, or any OpenAI-compatible API
• Markdown reports: Human-readable persistent knowledge base
• Universal notifications: Email, Discord, Slack, and more via Shoutrrr
• Single binary with no runtime dependencies

Target Audience

• DevOps engineers and system administrators
• Teams who need intelligent log analysis and anomaly detection

Quick Start

# Docker run
docker run --rm -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -v ./dlia-data:/data \
  -e DLIA_LLM_API_KEY=your-key-here \
  -e DLIA_LLM_MODEL=gpt-4o-mini \
  zorak1103/dlia:latest scan

Core Advantages

• LLM-powered semantic log analysis
• Automatic anomaly detection and trend tracking
• Privacy-first with automatic data anonymization
• Single binary with native Docker integration

System Requirements

• OS: Cross-platform (via Docker)
• Additional Dependencies: Docker, LLM API access