DriftSiren

Real‑Time Data Drift & Data Quality Monitoring Platform

A production‑style, full‑stack platform to detect data drift, monitor data quality, and alert in real time.

Key Features

• Agent SDK sends datasets/events to the backend securely.
• Real‑time drift metrics (PSI, KS, Chi‑Squared) computed in background jobs.
• Live alert stream via WebSocket; historical metrics stored in Postgres.
• Clean dashboard with charts (Recharts) and filters.
• Role‑based users (stub/JWT) and organizations (demo scope).
• Docker Compose one‑command startup. K8s manifests included for cluster deployments.

Project Structure

• https://github.com/AmSh4/DriftSiren/tree/main/Structure

Quick Start (Docker)

1. Clone the repo and create .env from .env.example.
2. Run: docker compose up --build.
3. Open http://localhost:3000 (frontend) and http://localhost:8000/docs (backend API).

Demo creds: any email works; click "Sign In (Demo)" to get a JWT.

Sample Run (without Docker)

• Backend: cd backend && uvicorn app.main:app --reload
• Worker: cd backend && celery -A app.celery_app.app worker --loglevel=INFO
• Frontend: cd frontend && npm i && npm run dev

Security Notes

Replace SECRET_KEY, add proper auth, HTTPS, and harden CORS before production.