❓ Frequently Asked Questions

FinOpsMetrics is an enterprise FinOps platform for AI/ML cost observability and optimization. It provides real-time cost monitoring, LLM-powered recommendations, and multi-cloud tracking (AWS, Azure, GCP) to help teams potentially reduce infrastructure costs.

Most teams are up and running in under 5 minutes. Simply install via pip (pip install finopsmetrics), configure your cloud credentials, and launch the dashboard. The system automatically discovers resources and starts tracking costs immediately.

Yes! FinOpsMetrics is 100% free under the Apache 2.0 license. There are no hidden costs, no premium tiers, and no credit card required. You can use all features without any limitations.

Teams may achieve potential cost savings on AI/ML infrastructure using FinOpsMetrics through:
• GPU right-sizing
• Auto-scaling policies
• Spot instance optimization
• LLM-powered recommendations

Actual savings vary based on workload characteristics, current infrastructure setup, and optimization implementation.

Spot instances are unused cloud capacity available at significant discounts (often up to 90%) compared to on-demand pricing. FinOpsMetrics identifies fault-tolerant workloads suitable for spot instances and provides specific recommendations. Actual discounts vary by region, availability zone, and instance type.

FinOpsMetrics optimizes costs for NVIDIA A100 (80GB), H100 (80GB), L4, T4, and V100 GPUs across AWS, GCP, and Azure. It provides specific recommendations like:
• "Consider switching from A100 to L4 for inference workloads"
• "Evaluate spot instances for fault-tolerant training jobs"
• "Right-size GPU allocation based on actual utilization metrics"

FinOpsMetrics v0.3.0 is the latest release, featuring:
• Plugin Architecture - 7 plugin types for unlimited extensibility
• Persistence Layer - 4 storage backends (In-Memory, SQLite, PostgreSQL, TimescaleDB)
• 7 Telemetry Agents - AWS, Azure, GCP, Databricks, Snowflake, SaaS, Generic
• 8 SaaS Integrations - MongoDB Atlas, Redis Cloud, GitHub Actions, DataDog, Kafka, Vercel, Docker Hub, Elasticsearch
• Persona-Specific Insights - Tailored insights for CFO, Engineers, Finance Analysts, and Business Leaders

Install via: pip install finopsmetrics

FinOpsMetrics supports:
• AWS - 50+ services including EC2, S3, Lambda, SageMaker, RDS, EKS
• Azure - 45+ services including VMs, Storage, Azure ML, Functions, AKS
• GCP - 40+ services including Compute Engine, Storage, Vertex AI, BigQuery, GKE
• Databricks - Cluster costs, DBU tracking, job attribution
• Snowflake - Warehouse costs, query attribution, storage tracking

You can monitor costs across all providers in a single unified dashboard.

FinOpsMetrics can integrate with OpenAI (GPT-4), Anthropic (Claude), Azure OpenAI, or Ollama (local models) to analyze your infrastructure and provide intelligent optimization recommendations. It analyzes your usage patterns, workload types, and costs to suggest specific actions like GPU downsizing, auto-scaling policies, or spot instance usage. You can also use rule-based recommendations without an LLM.

Yes! FinOpsMetrics tracks costs for Kubernetes clusters (EKS, GKE, AKS) including pod-level cost attribution, node right-sizing recommendations, and cluster auto-scaling policies. It helps optimize container workloads and provides namespace-level cost breakdowns.

Yes! FinOpsMetrics tracks LLM API usage and costs for OpenAI (GPT-4, GPT-3.5), Anthropic (Claude), and other providers. It monitors token usage, calculates costs per request, and provides optimization recommendations like caching, prompt optimization, or model selection.

FinOpsMetrics provides flexible cost attribution by team, project, environment, or custom tags. You can track costs by ML model, training run, department, or any custom dimension. The system automatically aggregates costs across services and provides detailed reports for chargeback or showback.

Yes! FinOpsMetrics v0.3.0 features a comprehensive plugin architecture with 7 plugin types:
• Telemetry Plugins - Custom data collectors for any cloud provider or service
• Attribution Plugins - Custom cost allocation logic
• Recommendation Plugins - ML-powered optimization suggestions
• Dashboard Plugins - Custom visualization widgets
• Integration Plugins - Connect to external tools (JIRA, ServiceNow, etc.)
• Notification Plugins - Custom alert channels
• Policy Plugins - Governance and compliance rules

Plugins can be distributed via PyPI and discovered through the plugin marketplace. See the Plugin Architecture documentation for details.

FinOpsMetrics v0.3.0 includes 8 fully functional SaaS integrations:
• MongoDB Atlas - Cluster costs (M10-M300 tiers), replication, sharding
• Redis Cloud - Database instances, throughput, multi-cloud support
• GitHub Actions - Workflow minutes, per-OS pricing, storage costs
• DataDog - Host count, custom metrics, monitoring costs
• Confluent Kafka - Cluster costs, CKU tracking, all tiers
• Vercel - Projects, deployments, bandwidth, function invocations
• Docker Hub - Repositories, pull counts, private repo costs
• Elasticsearch Cloud - Deployments, memory usage, data transfer

All integrations are production-ready and can be configured in the SaaS Services Agent.

FinOpsMetrics supports 4 storage backends (since v0.2.1):
• In-Memory - Ultra-fast, for development/testing, ephemeral
• SQLite - File-based, zero configuration, perfect for small deployments
• PostgreSQL - Production-grade, ACID compliance, enterprise ready
• TimescaleDB - Time-series optimized, compression, retention policies, ideal for large-scale monitoring

Choose the backend that fits your scale and requirements. You can configure retention periods (90 days, 1 year, custom) to manage storage costs.

FinOpsMetrics provides tailored insights for different organizational roles:
• CFO - Unit economics, ROI analysis, budget variance, cost forecasting
• Engineers - Idle resources, cost spikes, rightsizing opportunities, ARM migration
• Finance Analysts - Forecast variance, trend analysis, cost breakdowns, QoQ comparisons
• Business Leaders - Growth efficiency, capacity headroom, revenue-to-infrastructure ROI

Each persona receives insights optimized for their decision-making needs, delivered through configured channels (Slack, Email, Teams).

FinOpsMetrics uses 7 intelligent telemetry agents that automatically discover and track costs:
• AWS Agent - Integrates with Cost Explorer API and CloudWatch
• Azure Agent - Uses Cost Management API and Azure Monitor
• GCP Agent - Leverages Cloud Billing API and Cloud Monitoring
• Databricks Agent - Tracks cluster costs and DBU consumption
• Snowflake Agent - Monitors warehouse costs and query attribution
• SaaS Services Agent - Collects costs from 8 SaaS platforms
• Generic Agent - Extensible for custom/on-premise infrastructure

Deploy once and agents continuously discover new resources and track costs with minimal overhead.

FinOpsMetrics offers multiple support channels:
• Email Support: Contact our team at finops@infinidatum.net
• Documentation: Comprehensive guides and tutorials
• API Reference: Detailed API documentation
• Knowledge Base: FAQ section with common questions

Response Time: Our support team typically responds within 24 hours during business hours.

Absolutely! FinOpsMetrics is designed for production environments with features like budget alerts, cost attribution, role-based dashboards (CFO, COO, Infrastructure), and real-time monitoring. It's built to handle large-scale AI/ML workloads.

FinOpsMetrics is 100% free with no vendor lock-in. Unlike commercial tools that charge per-node or percentage of spend, FinOpsMetrics has zero cost. It's specifically designed for AI/ML workloads with GPU optimization, LLM cost tracking, and intelligent recommendations. You have full control over your data and can self-host or customize the platform.