Multi-cloud Cost Management Platform with FOCUS
Talkat Open Source Summit Japan + AI_dev: Open Source GenAI & ML Summit + Automotive Linux Summit 2025 · 2025-12-08..12-10·
summary
Service-oriented Multi-Cloud Cost Management Platform normalizes AWS, Azure and other billing exports into the FOCUS schema, stores raw exports in S3/Blob, transforms with Glue/ADF into a catalog, visualizes in AWS Managed Grafana, and layers MLOps for anomaly detection using SageMaker and an XGBoost model. Ingest runs under a minute for small daily volumes (AWS ~1 min, ADF 47s). Anomaly model reports AUC 0.885, precision 0.97 on three months of data.
AI-written summary · about
chapters (13)
- 0:00Multi-cloud cost challenges
- 1:51Cost tracking pain points
- 4:02Normalizing costs with FOCUS
- 6:54Platform architecture overview
- 8:48AWS and Azure pipelines
- 10:56FOCUS mapping differences
- 12:24Cost anomaly detection
- 14:10Ingestion and normalization performance
- 15:37Visualizing normalized costs
- 17:50Validating ML anomalies
- 18:56FOCUS and SOA benefits
- 20:01Cost versus business value
- 21:55Q&A
related talks
17mOptimizing Multi-Cloud Platforms with AI Ops and FinOps
36mFinOps & Observability: technical implementation
46mWhen the CFO Comes Knocking: Our Journey to a Cost Efficient Data Platform
46mCloud cost chaos? Get clarity with FinOps cloud cost management
27mWith FinOps Through the Cloud Cost Jungle
20mCloud Cost Management Updates
34mOptimizing Cloud Efficiency with Prompt Engineering
19mFinOps for business results