Telemetry Engine
An ultra-lightweight system performance monitor written in pure Python.
⚡ Active Tier 1 Project (Pre-release).
~ About the Engine
Instead of relying entirely on psutil or spawning subprocesses, this engine hooks directly into kernel data structures (/proc/stat and /proc/meminfo) on Linux to capture real-time CPU and memory statistics with an almost undetectable resource footprint.
v0.2.3 Update: The engine now includes a cross-platform hardware abstraction layer (HAL). If /proc is missing (like on Windows or macOS), it seamlessly falls back to psutil. It also introduces battery telemetry, tracking charge capacity, power draw, and hardware degradation.
~ Tech Stack & Architecture
~ The Telemetry Series
Building this engine meant learning how operating systems expose their internal metrics. I documented every roadblock, deep dive, and "gotcha" I encountered along the way.
Why Linux shows 95% RAM usage when idle, and how to parse kernel data using pure Python.
Cross-platform telemetry, the cpu_percent interval gotcha, and building a fallback system.
Talking to the Embedded Controller via /sys and handling desktop fallbacks.
Currently researching how to extract thermal zones without external libraries.
~ How to Run
Prerequisites: Python 3.10+ and psutil (for cross-platform fallback support).
1. Clone and set up the environment:
git clone https://github.com/Tharana-Dev/telemetry-engine.git
cd telemetry-engine
python3 -m venv .venv
# Activate the environment:
source .venv/bin/activate # Linux / macOS
# .venv\Scripts\activate # Windows
pip install -r requirements-dev.txt2. Run the live dashboard:
python3 main.py3. Run the test suite:
pytest -v~ Releases
Download the v0.2.3 Pre-release directly, or visit GitHub to see the latest version and changelogs.