Drift-SRE

Real-time Anomaly Detection Library for SRE

Drift-SRE is a lightweight, embeddable Python library that monitors server metrics and detects anomalies in real-time, with intelligent Discord notifications. Perfect for integrating into your existing applications without requiring separate services.

Quick Start

from drift import DriftMonitor

# Initialize with Discord webhook
monitor = DriftMonitor(
    discord_webhook="https://discord.com/api/webhooks/YOUR_WEBHOOK_URL",
    check_interval=5  # seconds
)

# Customize specific metrics (optional)
monitor.configure_metric('ram_percent', threshold=10.0, drift=2.0)
monitor.configure_metric('cpu_percent', threshold=25.0, drift=5.0)

# Start monitoring in background
monitor.start()

# That's it! Monitor runs in background thread

Installation

pip install drift-sre

Features

• Embeddable: Runs in the same process as your application
• Zero External Services: No separate monitoring server required (except Discord for notifications)
• Smart Defaults: Sensible configurations for all system metrics out of the box
• Per-Metric Tuning: Configure each metric independently with CUMSUM or EWMA algorithms
• Discord Notifications: Beautiful embeds with rate limiting and recovery notifications
• Custom Metrics: Register your own metrics with custom collectors
• Thread-Safe: Safe to use with Flask, FastAPI, Django, and other frameworks

Monitored Metrics

• CPU usage percentage
• RAM usage percentage
• Disk read/write throughput
• Network sent/received bytes
• System load average
• Active network connections

Algorithms

CUMSUM (Cumulative Sum)

Best for stable metrics like CPU and RAM. Detects sustained shifts from normal behavior.

EWMA (Exponentially Weighted Moving Average)

Best for variable metrics like network traffic and disk I/O. Adapts to changing patterns.

Documentation

• Getting Started - Installation and basic setup
• Configuration Guide - Tuning detection parameters
• Metrics Guide - Understanding each metric
• API Reference - Complete API documentation
• Advanced Usage - Custom metrics and integrations
• Troubleshooting - Common issues and solutions

Examples

See the examples directory for:

• Basic usage
• Flask integration
• FastAPI integration
• Custom metrics

Requirements

• Python 3.8+
• psutil (for system metrics)
• requests (for Discord webhooks)

License

MIT License

Contributing

Contributions welcome! Please see our contributing guidelines.