FinOps Hub
FinOps Overview
The StackFlow FinOps Hub provides a comprehensive view of cloud spending across all connected accounts. It ingests AWS Cost and Usage Reports (CUR), Azure Cost Management data, and GCP Billing exports to provide a unified multi-cloud cost view. The hub integrates with the CMDB to attribute costs to business services, departments, and teams.
- Cost Explorer: Enabled with tag-based cost allocation; tags
StackFlowServiceandStackFlowTenantapplied to all resources - S3 CUR: Cost and Usage Report delivery to
stackflow-cur-373544523367bucket in Parquet format, daily granularity - IAM:
StackFlowFinOpsRolewithce:*,s3:GetObjecton CUR bucket - DynamoDB:
StackFlow_BudgetAlerttable for cost anomaly and budget threshold alerts
Cost Attribution
Cost attribution maps cloud spending to organizational dimensions using tags. Define tagging policies in FinOps Hub → Attribution → Tag Policies. Resources that conform to the tag policy are automatically attributed to the correct department and team. Untagged resources are shown in the "Unattributed" bucket and trigger a compliance finding.
| Required Tag | Purpose | Example Values |
|---|---|---|
Environment | Production vs. non-production cost separation | prod, staging, dev |
Team | Team-level cost attribution | platform-eng, security, data-eng |
Service | Business service attribution | stackflow-api, auth-service, data-pipeline |
CostCenter | Finance system integration | CC-1042, CC-2034 |
Budget Management
Set monthly, quarterly, and annual budgets by account, service, team, or cost center. Budget alerts fire at configurable thresholds (default: 80% and 100% of budget) via the stackflow-security-findings SNS topic (or a dedicated budget alerts topic). Budget overruns beyond 110% auto-create a P2 incident for the Cloud team.
Forecasting
The FinOps forecast engine uses 90-day historical spend patterns and seasonal adjustments to project month-end and quarter-end costs. Forecasts are displayed as confidence intervals (P10/P50/P90) to communicate uncertainty. The AI engine supplements statistical forecasting with contextual factors like planned infrastructure changes extracted from the Change Management module.
Anomaly Detection
FinOps anomaly detection continuously monitors spending for unusual patterns — unexpected spikes, sudden drops (potential issue with billing ingestion), or gradual drift above baseline. Detected anomalies are flagged in the FinOps dashboard and generate StackFlow notifications. The AI engine provides a natural language explanation of each anomaly and links to the responsible resources in the CMDB.
aws ce get-anomalies --anomaly-monitor-arn arn:aws:ce::373544523367:anomalymonitor/stackflow-monitor --date-interval '{"StartDate": "2026-05-01", "EndDate": "2026-05-18"}' --region us-east-1