-
Notifications
You must be signed in to change notification settings - Fork 0
Home
SentraCore is a Local System Behavior Intelligence Platform. Unlike traditional system monitors that display raw metrics, SentraCore continuously interprets system performance, detects anomalies against a learned baseline, explains the root cause of slowdowns, and generates predictive warnings before critical resource exhaustion occurs.
To shift system monitoring from reactive ("What broke?") to proactive and explainable ("Why is it breaking, and when will it become critical?").
This is achieved through a six-phase intelligence pipeline that transforms raw telemetry into actionable, human-readable insights.
SentraCore doesn't rely on hardcoded thresholds. Instead, it builds a statistical baseline of your specific machine's behavior (taking into account the time of day). The Anomaly Detector then calculates Z-Scores to alert you only when your system behaves abnormally compared to its historical patterns.
When SentraCore detects a high-stress state, it doesn't just ring an alarm. The RCA Engine cross-references the bottleneck (e.g., Memory) with recent system events and running processes to generate a clear, human-readable explanation:
"System is experiencing critical memory pressure. The process 'chrome.exe' (PID 1234) is the primary suspect."
Moving beyond current state analysis, SentraCore uses Exponential Moving Averages (EMA) to forecast future states. It generates:
- A Degradation Risk Score (0-100%).
- Time-to-Exhaustion (ETA) countdowns (e.g., predicting that Memory will hit 98% saturation in 45 seconds if current trends continue).
The traditional single-dimensional "Stress Score" is replaced by a holistic System Stability Index (1–100). A score of 100 represents perfect health. Penalties are applied dynamically based on instantaneous resource stress, forward-looking predictive risk, and statistical anomaly deviation.
SentraCore is structured as two decoupled layers:
1. Headless Engine (Python)
- Collects system telemetry via
psutil. - Processes data through the intelligence pipeline.
- Exposes a REST API (
FastAPI) and a live WebSocket stream. - Runs silently in the background as a Windows startup process.
2. Intelligence Dashboard (Flutter)
- A responsive desktop application built with Dart and Flutter.
- Subscribes to the engine's WebSocket for real-time updates.
- Visualizes the Stability Index, predictive ETAs, Root Cause Analysis data, and historical metric charts.
To install SentraCore on your Windows machine, visit the Releases page and download the latest SentraCore_Setup.exe.
For contributors and developers, refer to the following guides: