School Nutrition Platform: 6 Microservices with Grafana Loki Observability and Auto-Scaling for Peak Lunch Hours

Why This Matters

These results demonstrate the tangible business value of investing in the right technology infrastructure — from improved reliability to measurable cost savings.

MealManage's Challenge

MealManage provides school meal ordering, parent notifications, and caterer management for school districts across the country. Their monolithic application had grown organically, and as they expanded to more districts, deployment bottlenecks were slowing down feature delivery.

Operational challenges:

• Monolithic deployment meant any update required full system deployment during school hours, risking lunch ordering disruptions
• No visibility into application logs across different components — debugging took hours of SSH sessions
• Manual deployments slowing down feature releases — teams waited weeks for deployment windows
• Peak lunch hours (11am-1pm) required manual scaling intervention to handle order surges
• Caterer partners needed dedicated APIs but shared codebase made changes risky

Implementation Details

• Web Service — Parent-facing ordering portal
• API Service — Core business logic and data access
• Email Service — Notification delivery and templating
• Partner API — Caterer inventory and fulfillment integration
• Parent API — Mobile app backend
• Caterer Web — Kitchen management dashboard

Centralized Observability

• Built centralized logging with Grafana + Loki using Fluent Bit sidecars via AWS FireLens
• Created dashboards for order funnel metrics, service latencies, and error rates

Auto-Scaling & CI/CD

• Configured auto-scaling at 75% CPU with fast cooldowns for lunch rush handling
• Created 6 independent CI/CD pipelines with AWS CodePipeline for autonomous deployments
• Implemented host-based routing on ALB for service isolation