Cost Calculations Guide

Deep dive into how FinFocus Core calculates projected and actual infrastructure costs.

Overview
Projected Cost Calculations
Actual Cost Calculations
Cost Methodologies
Data Sources
Aggregation and Grouping
Accuracy and Limitations
Best Practices

Overview

FinFocus Core provides two types of cost analysis:

Projected Costs: Estimates based on resource specifications and pricing data
Actual Costs: Historical spending retrieved from cloud provider APIs

Understanding how these calculations work helps you interpret results and optimize your infrastructure costs.

Projected Cost Calculations

Projected costs estimate monthly spending based on resource configurations and pricing specifications.

Calculation Flow

graph TD
    A[Pulumi Plan JSON] --> B[Resource Extraction]
    B --> C[Resource Classification]
    C --> D{Plugin Available?}
    D -->|Yes| E[Plugin Pricing Query]
    D -->|No| F[Local Spec Lookup]
    E --> G[Cost Calculation]
    F --> G
    G --> H[Monthly Projection]
    H --> I[Result Aggregation]

Resource Classification

Resources are classified by type and provider:

{
  "type": "aws:ec2/instance:Instance",
  "provider": "aws",
  "service": "ec2",
  "resourceType": "instance"
}

Pricing Sources (Priority Order)

Plugin Data: Live pricing from cloud provider APIs
Local Specs: YAML/JSON pricing specifications
Fallback Estimates: Default values by resource category

Monthly Calculation Formula

Monthly Cost = Hourly Rate × Hours Per Month (730)

Constants used:

Hours per month: 730 (average accounting for varying month lengths)
Hours per day: 24
Days per month: 30.42 (average)

Example Calculation

AWS EC2 t3.micro instance:

# Pricing spec
pricing:
  onDemandHourly: 0.0104

# Calculation
Monthly Cost = 0.0104 × 730 = $7.59

Resource-Specific Calculations

Compute Resources

EC2 Instances: Based on instance type, region, and pricing model
ECS Tasks: vCPU and memory allocation with Fargate pricing
Lambda Functions: Execution time and memory allocation

Storage Resources

S3 Buckets: Storage class, region, and estimated data volume
EBS Volumes: Volume type, size, and IOPS provisioning
EFS File Systems: Storage class and throughput mode

Database Resources

RDS Instances: Engine type, instance class, and storage
DynamoDB Tables: Provisioned vs on-demand, read/write capacity
DocumentDB Clusters: Instance type and storage requirements

Network Resources

Load Balancers: Type (ALB/NLB/CLB) and estimated data transfer
NAT Gateways: Data processing charges
VPC Endpoints: Interface vs Gateway endpoints

Fallback Pricing

When specific pricing is unavailable, default estimates are used:

const (
    defaultDatabaseMonthlyCost = 50.0   // Database resources
    defaultStorageMonthlyCost = 5.0     // Storage resources
    defaultComputeMonthlyCost = 20.0    // Compute resources
)

Actual Cost Calculations

Actual costs retrieve historical spending data from cloud provider billing APIs through plugins.

Data Retrieval Flow

graph TD
    A[Date Range Request] --> B[Plugin Selection]
    B --> C[API Authentication]
    C --> D[Cost Data Query]
    D --> E[Resource Matching]
    E --> F[Data Aggregation]
    F --> G[Time Series Processing]
    G --> H[Result Formatting]

Time Range Processing

Date Format Support

Simple dates: 2025-01-01 (assumes start/end of day)
RFC3339: 2025-01-01T14:30:00Z (precise timestamps)
Partial timestamps: 2025-01-01T14:30:00Z (timezone aware)

Time Zone Handling

All times converted to UTC for consistency
Plugin APIs may return data in different time zones
Results always include timezone information

Resource Matching

Resources from Pulumi plans are matched to cost data using:

Resource ARN/ID: Direct matching where available
Resource Tags: Tag-based identification
Resource Type: Service and type matching
Fuzzy Matching: Heuristic matching based on properties

Aggregation Methods

By Time Period

{
  "dailyCosts": [12.45, 13.67, 11.23, ...],
  "totalCost": 156.78,
  "costPeriod": "2025-01-01 to 2025-01-31"
}

By Resource

{
  "resourceType": "aws:ec2/instance:Instance",
  "resourceId": "web-server",
  "totalCost": 89.45
}

By Service/Provider

{
  "summary": {
    "byProvider": { "aws": 156.78 },
    "byService": { "ec2": 89.45, "s3": 23.12, "rds": 44.21 }
  }
}

Cost Methodologies

Projected vs Actual Cost Differences

Aspect	Projected Costs	Actual Costs
Data Source	Pricing specifications	Billing APIs
Timing	Pre-deployment	Post-deployment
Accuracy	Estimates	Actual spending
Variability	Static pricing	Usage-based fluctuations
Scope	Resource configurations	Realized consumption

Common Discrepancies

Usage Patterns

Projected: Assumes 100% uptime (730 hours/month)
Actual: Reflects real usage patterns, scaling, downtime

Pricing Models

Projected: Often uses on-demand pricing
Actual: Includes reserved instances, spot instances, savings plans

Additional Services

Projected: Only explicitly defined resources
Actual: Includes data transfer, API calls, storage I/O

Accuracy Factors

High Accuracy Scenarios

Well-defined resource specifications
Stable usage patterns
On-demand pricing models
Simple resource configurations

Lower Accuracy Scenarios

Auto-scaling configurations
Complex data transfer patterns
Reserved instance utilization
Multi-service dependencies

Data Sources

Plugin-Based Data Sources

Kubecost Plugin

Source: Kubecost API
Coverage: Kubernetes workloads, container costs
Granularity: Pod, namespace, cluster level
Update Frequency: Near real-time

AWS Cost Explorer Plugin

Source: AWS Cost Explorer API
Coverage: All AWS services
Granularity: Resource, service, account level
Update Frequency: Daily (24-hour delay)

Cloud Provider Plugins

Azure Cost Management: Azure spending data
GCP Cloud Billing: Google Cloud costs
Multi-cloud: Aggregated cross-cloud costs

Local Pricing Specifications

Format Example

provider: aws
service: ec2
sku: t3.micro
currency: USD
pricing:
  instanceType: t3.micro
  onDemandHourly: 0.0104
  reservedHourly: 0.0074
  spotHourly: 0.0031
  monthlyEstimate: 7.59
metadata:
  region: us-west-2
  operatingSystem: linux
  tenancy: shared
  vcpu: 2
  memory: 1

Spec Discovery

Check ~/.finfocus/specs/ directory
Match by provider, service, and SKU
Fallback to generic service pricing
Use default estimates if no match found

Aggregation and Grouping

Grouping Options

By Resource (`group-by resource`)

{
  "resourceType": "aws:ec2/instance:Instance",
  "resourceId": "web-server-1",
  "totalCost": 45.67
}

By Type (`group-by type`)

{
  "groupKey": "aws:ec2/instance:Instance",
  "totalCost": 134.56,
  "resourceCount": 3
}

By Provider (`group-by provider`)

{
  "groupKey": "aws",
  "totalCost": 456.78,
  "services": ["ec2", "s3", "rds"]
}

By Date (`group-by date`)

{
  "date": "2025-01-15",
  "totalCost": 15.67,
  "resources": [...]
}

By Tags (`group-by tag:key=value`)

{
  "tagKey": "Environment",
  "tagValue": "production",
  "totalCost": 234.56,
  "matchedResources": 12
}

Summary Calculations

Cost Summary Structure

{
  "summary": {
    "totalMonthly": 456.78,
    "totalHourly": 0.626,
    "currency": "USD",
    "byProvider": { "aws": 456.78 },
    "byService": { "ec2": 234.56, "s3": 78.9, "rds": 143.32 },
    "byAdapter": { "kubecost": 234.56, "aws-spec": 222.22 }
  }
}

Aggregation Rules

Sum: Total costs across resources
Average: For rate-based metrics
Count: Number of resources in group
Currency: Unified to single currency (USD default)

Accuracy and Limitations

Projected Cost Limitations

Pricing Data Freshness

Local specs may be outdated
Regional pricing variations
Promotional pricing not reflected

Usage Assumptions

Assumes consistent 24/7 operation
No auto-scaling considerations
Static resource configurations

Missing Costs

Data transfer charges
API request costs
Third-party service integrations

Actual Cost Limitations

Data Lag

Billing APIs have 12-48 hour delays
Real-time costs not available
Processing time for complex queries

Attribution Challenges

Resource tagging inconsistencies
Shared resource costs
Multi-account cost allocation

Plugin Dependencies

Requires plugin installation and configuration
API rate limits and authentication
Network connectivity requirements

Improving Accuracy

For Projected Costs

Keep specs updated: Regularly update pricing specifications
Use plugins: Leverage live pricing APIs when available
Include usage patterns: Account for auto-scaling and variable usage
Regional awareness: Use region-specific pricing

For Actual Costs

Consistent tagging: Implement comprehensive resource tagging
Regular reconciliation: Compare projected vs actual regularly
Multiple data sources: Use multiple plugins for validation
Time range considerations: Allow for billing data processing delays

Best Practices

Cost Estimation Workflow

Generate Projections: Start with projected cost estimates
Deploy and Monitor: Track actual costs post-deployment
Compare and Analyze: Identify discrepancies and patterns
Optimize and Iterate: Adjust resources based on findings

Tagging Strategy

Implement consistent tagging for accurate cost attribution:

// Example Pulumi resource with comprehensive tags
const webServer = new aws.ec2.Instance('web-server', {
  instanceType: 't3.micro',
  tags: {
    Environment: 'production',
    Team: 'backend',
    Project: 'web-app',
    CostCenter: 'engineering',
    Application: 'api-server',
    Owner: 'team@company.com',
  },
});

Regular Validation

# Monthly cost validation workflow
finfocus cost projected --pulumi-json plan.json --output json > projected.json
finfocus cost actual --pulumi-json plan.json --from $(date -d '1 month ago' +%Y-%m-01) --output json > actual.json

# Compare results
jq -r '.summary.totalMonthly' projected.json
jq -r '.summary.totalMonthly' actual.json

Cost Optimization

Identify High-Cost Resources

# Find most expensive resources
finfocus cost actual --pulumi-json plan.json --from 2025-01-01 --output json | \
  jq '.resources | sort_by(.totalCost) | reverse | .[0:5]'

Track Cost Trends

# Generate daily cost trend
for day in {01..31}; do
  finfocus cost actual --pulumi-json plan.json --from 2025-01-$day --to 2025-01-$day --output json | \
    jq -r "[\"2025-01-$day\", .summary.totalMonthly] | @csv"
done > cost-trend.csv

User Guide - Complete usage instructions
Plugin System - Plugin development and configuration
Troubleshooting - Common issues and solutions
Examples - Sample pricing specs and configurations