SignalGuard Observability Platform

🔍 Overview

SignalGuard is an end-to-end observability demo environment designed to showcase:

Production-grade monitoring & observability practices
Instrumentation of microservices using Prometheus metrics
A separate Python anomaly detection service that behaves like an early-stage AIOps layer
Grafana dashboards for system and service health
A modern web dashboard (GitHub Pages) with live-updating charts and simulated incident patterns

This project demonstrates how monitoring, metrics pipelines, anomaly detection, and dashboards can be composed into a coherent observability platform—the same architectural pattern used in real SRE, Ops, and AIOps workflows.

🚀 Features

🧩 Synthetic Microservices (FastAPI)

Latency histograms
Error rate spikes
Request throughput
Business metrics
Background load generation

All exported as Prometheus metrics.

🤖 Anomaly Detection Service

Queries Prometheus ranges (via HTTP API)
Computes statistical anomaly score
Emits signalguard_anomaly_flag
Emits signalguard_anomaly_score

Models a lightweight AIOps system.

📈 Prometheus & Grafana

Auto-provisioned using Docker
Prometheus scrapes all services
Grafana dashboard JSON loads automatically
Contains latency p95, error rate, throughput
Anomaly score panels

💠 Standalone GitHub Pages Dashboard

Completely static
NO backend required
Runs on simulated data
Beautiful glassmorphism UI
Real-time charts
Synthetic anomalies

Ideal for interviews or recruiters.

🏗️ Architecture

signalguard-observability ├── services/ │ ├── app/ # FastAPI microservice │ │ ├── app.py # Routes + Prometheus metrics + synthetic load │ │ ├── static/ # Frontend assets (for local dashboard) │ │ └── Dockerfile │ │ │ ├── anomaly-detector/ # Python anomaly detection service │ ├── detector.py # Pulls metrics from Prometheus │ └── Dockerfile │ ├── observability/ │ ├── prometheus/ # Scrape config & rules │ └── grafana/ # Provisioning + dashboards │ ├── docs/ # GitHub Pages dashboard (simulated) │ ├── index.html │ ├── styles.css │ └── script.js │ ├── docker-compose.yml └── README.md

🛠️ Running the Full System (Local Deployment)

Requirements:

Docker
Docker Compose

🔧 Step 1 — Clone Repo

git clone https://github.com/danial-amin/signalguard-observability

🔧 Step 2 — Run the stack

docker-compose build
docker-compose up

🌐 Services

Service	Description	URL
FastAPI Microservice	Synthetic endpoints + metrics	http://localhost:8000
FastAPI Local Dashboard	Interactive dashboard	http://localhost:8000/dashboard
Prometheus	Metrics explorer	http://localhost:9090
Grafana	Dashboards (login: admin / admin)	http://localhost:3000
Anomaly Detector API	Exposes anomaly flags	http://localhost:8001

🎯 Why This Project Matters

This project demonstrates:

✔ Observability Architecture

Microservice instrumentation
Metrics pipelines
Prometheus scraping
Grafana provisioning
Real-time monitoring

✔ AIOps Foundations

Automated anomaly detection
Metric-driven incident detection
Flags + scoring + thresholds
Modular ML microservice pattern

✔ System Design Skills

Independent services
Clean dockerized environment
Clear separation of concerns
Realistic production-like pipeline

✔ Frontend + UX

Custom dashboard
Charts, alerts, anomaly cards
Modern, responsive UI

This makes you look both engineering-strong AND design-aware — a combination that companies like Kyndryl, Supercell, Samsung, and Oura love.

🚦 GitHub Pages Deployment (Static Demo)

This repo includes a docs/ folder that GitHub Pages serves automatically.

How it works:

Every 5 seconds, the dashboard generates synthetic request counts, error rates, and anomaly detection patterns
No backend required
Perfectly safe to load anywhere

🚀 View Live Dashboard Demo

📜 License

MIT License

🤝 Contributions

Open to pull requests, improvements, and extensions—especially around:

Additional anomaly detection techniques
Predictive models
More microservices
Histogram-based ML inputs