System Architecture — 3 Layers
NODES
🖥️
AWS Instances
eu-north · eu-west
🖥️
Hetzner Nodes
Dedicated Servers
🖥️
Service Nodes
Application Hosts
▲
FastAPI :880 → JSON metrics (CPU, RAM, Net, Uptime, Traffic)
▼
COLLECT
📥
Data Collectors
Multiprocessing Pool
📄
CSV Store
Per-Provider Files
CSV Read · 0.5s Refresh
▼
DISPLAY
🖥️
Terminal TUI
Curses · Color-coded
🌐
Web Dashboard
PHP · Long-Polling
🔔
Alert Engine
Threshold Rules
Key Features
⚡
Parallel Data Collection
Multiprocessing Pool fetches metrics from all servers simultaneously. Scales to 30+ nodes with no bottleneck.
🖥️
Curses Terminal UI
Full-screen ncurses dashboard with color-coded health indicators, scrollable sections, and 0.5s auto-refresh.
📊
Multi-Provider View
Unified view across AWS, Hetzner, and custom nodes — each with provider-specific health thresholds and logic.
🚨
Smart Anomaly Rules
Context-aware alerts: high CPU + low traffic = problem. High traffic + high CPU = normal. No false positives.
🔄
FastAPI Resource Agent
Async agent on each node exposing CPU, RAM, network speed, traffic, and service uptime via REST endpoint.
🛡️
Auto-Healing Triggers
Automatic service restarts when RAM exceeds 96%. Service management endpoints for remote reboot and restart.
Tech Stack
Python
FastAPI
asyncio
Multiprocessing
ncurses
Pandas
psutil
uvicorn
REST API
CSV Pipeline
Subprocess Mgmt
Telegram Alerts
System Scale