Performance Monitoring with Grafana

Project

This project showcases a Dockerized service for serving a regression model and monitoring its performance in production. The core components include:

  • A Flask REST API that serves model predictions through endpoints.
  • A monitoring system that generates new sample data and logs data drift and concept drift metrics.
  • A Grafana dashboard for visualizing these metrics in real-time.
  • An alert pipeline that triggers notifications to a Discord channel whenever monitored metrics exceed a threshold value.
  • A containerized application (Docker and Docker Compose) for consistent deployment.

Grafana Dashboard

Discord Alerts

Code