Case Study

FinOps Cost Optimization Portal

Analytics portal that tracks cloud spend, flags anomalies and recommends cost saving actions.

PythonFastAPIPostgreSQLRedisDocker

Problem Statement

Engineering teams lacked visibility into cost spikes and underutilized resources across cloud services.

Dataset / Scale

Daily billing exports, 12+ services, multi-account aggregation with trend analysis.

Architecture Diagram

ETL layer transforms billing exports
Anomaly detection pipeline + rule-based alerts
Dashboard API with cached aggregates
Role based access for team reporting

ML / LLM Pipeline

Extract billing data → Normalize → AggregateDetect anomalies → Notify → ReportSurface recommendations → Track savings

System Design

  • FastAPI backend with async endpoints
  • Redis caching for dashboards
  • PostgreSQL analytics tables
  • Dockerized deployment

Engineering Challenges

  • Reducing noisy alerts for short lived spikes
  • Keeping dashboards under 200ms p95
  • Aligning cost tags across teams

Metrics

  • 25% faster anomaly detection
  • 18% monthly savings identified
  • 200ms p95 dashboard latency

Screenshots

Cost anomaly heatmap by service
Optimization recommendations view
Monthly savings report dashboard

Links