Agent Skills: Infrastructure Cost Estimation Methodology

Infrastructure Cost Estimation Methodology

CRITICAL REQUIREMENT: EXCLUDE ALL CREDITS BY DEFAULT

MANDATORY: Always exclude credits, promotional offers, and discounts to get TRUE infrastructure cost. Report:

1. Gross Cost (without credits) - PRIMARY NUMBER
2. Net Cost (with credits) - comparison only

Credits are temporary and mask real costs. Always plan for post-credit expenses.

Core Principles

1. Native Cost Tools First: Use cloud provider billing tools as primary source
2. Credits Excluded: Always exclude credits unless analyzing discount impact
3. Comprehensive Discovery: Identify ALL infrastructure components
4. Current Pricing: Research real-time standard pricing only
5. Python Calculations: Use Python for ALL numeric operations

Phase 1: Native Cost Estimation

Time Period Requirements

CRITICAL: Always use the most recent complete months for analysis:

• Primary Analysis: Last 3 complete months (most recent data)
• Trend Analysis: Last 6 complete months (for patterns)
• Never use data older than 6 months unless specifically requested
• Always specify actual date ranges in your analysis
• Determine current date first, then calculate recent complete months

Credit Exclusion Steps - CLI Commands

CRITICAL: Use these exact CLI commands to exclude credits and get true infrastructure costs:

AWS CLI Credit Exclusion

# Exclude all credits and promotional charges
aws ce get-cost-and-usage \
  --time-period Start=YYYY-MM-01,End=YYYY-MM-01 \
  --granularity MONTHLY \
  --metrics "BlendedCost" \
  --filter '{
    "Not": {
      "Dimensions": {
        "Key": "RECORD_TYPE",
        "Values": ["Credit", "Refund", "SavingsPlanNegation", "DiscountedUsage"]
      }
    }
  }' \
  --group-by Type=DIMENSION,Key=SERVICE

Azure CLI Credit Exclusion

Create query file exclude-credits.json:

{
  "type": "Usage",
  "timeframe": "Custom",
  "timePeriod": {
    "from": "YYYY-MM-01T00:00:00.000Z",
    "to": "YYYY-MM-01T00:00:00.000Z"
  },
  "dataset": {
    "granularity": "Monthly",
    "filter": {
      "not": {
        "dimensions": {
          "name": "ChargeType",
          "operator": "In",
          "values": ["Credit", "Refund", "RoundingAdjustment"]
        }
      }
    },
    "aggregation": {
      "totalCost": {
        "name": "PreTaxCost",
        "function": "Sum"
      }
    },
    "grouping": [{"type": "Dimension", "name": "ServiceName"}]
  }
}

Execute with:

az rest --method POST \
  --url "https://management.azure.com/subscriptions/$(az account show --query id -o tsv)/providers/Microsoft.CostManagement/query?api-version=2023-11-01" \
  --body @exclude-credits.json

Google Cloud Credit Exclusion

Note: Google Cloud requires BigQuery queries. No direct CLI credit filtering available.

# First, execute BigQuery to exclude promotional credits
bq query --use_legacy_sql=false '
SELECT 
  invoice.month,
  service.description,
  SUM(cost) as gross_cost_no_credits
FROM `PROJECT-ID.DATASET.gcp_billing_export_v1_BILLING-ACCOUNT-ID`
WHERE invoice.month IN ("YYYYMM", "YYYYMM", "YYYYMM")
GROUP BY invoice.month, service.description
ORDER BY gross_cost_no_credits DESC;'

Key Filter Parameters by Provider:

• AWS: Exclude RECORD_TYPE values: "Credit", "Refund", "SavingsPlanNegation", "DiscountedUsage"
• Azure: Exclude ChargeType values: "Credit", "Refund", "RoundingAdjustment"
• Google Cloud: Use BigQuery on billing export tables, no promotional credits included

Native Tools

AWS: Cost Explorer, Pricing Calculator, Budgets, Cost & Usage Reports, Billing Dashboard Azure: Cost Management + Billing, Pricing Calculator, Advisor, Resource Graph
Google Cloud: Cloud Billing, Pricing Calculator, Asset Inventory, Recommender

Phase 2: Resource Discovery

Methods: IaC analysis, live environment queries, API enumeration Categories: Compute, storage, networking, platform services, security, monitoring, development, data services Coverage: All environments, regions, scaling policies, shared resources

Phase 3: Pricing Research

Requirements:

• Official provider pricing pages only
• Region-specific standard rates
• No promotional or discount pricing
• Current pay-as-you-go rates

Phase 4: Python Calculations

Credit Analysis Template

def analyze_credit_impact(billing_data, analysis_period="recent_months"):
    """
    Analyze credit impact for recent complete months
    Calculate recent months dynamically based on current date
    """
    from datetime import datetime, timedelta
    
    # Determine current date and calculate recent complete months
    current_date = datetime.now()
    current_month = current_date.month
    current_year = current_date.year
    
    # Calculate the last 3 complete months
    recent_months = []
    for i in range(3):
        month_offset = i + 1
        if current_month - month_offset <= 0:
            month = 12 + (current_month - month_offset)
            year = current_year - 1
        else:
            month = current_month - month_offset
            year = current_year
        recent_months.append((year, month))
    
    recent_months.reverse()  # Put in chronological order
    
    # Filter billing data to recent complete months
    recent_billing_data = [
        charge for charge in billing_data 
        if (charge['date'].year, charge['date'].month) in recent_months
    ]
    
    analysis = {
        'analysis_period': f"Recent 3 complete months: {recent_months[0][1]}/{recent_months[0][0]} - {recent_months[2][1]}/{recent_months[2][0]}",
        'gross_monthly_cost': sum(
            charge['amount'] for charge in recent_billing_data 
            if charge['type'] in ['Usage', 'Tax', 'Fee']
        ) / 3,  # Average over 3 months
        'net_monthly_cost': sum(charge['amount'] for charge in recent_billing_data) / 3,
        'total_credits_applied': 0,
        'credit_sustainability': 'TEMPORARY - Assume all credits expire'
    }
    
    analysis['total_credits_applied'] = (
        analysis['gross_monthly_cost'] - analysis['net_monthly_cost']
    )
    
    return analysis

Basic Cost Calculation

def calculate_monthly_costs(resources, pricing_data):
    HOURS_PER_MONTH = 730
    total_cost = 0
    cost_breakdown = {}
    
    for service_name, service_config in resources.items():
        service_cost = 0
        
        # Fixed costs (standard hourly rates)
        if 'instances' in service_config:
            hourly_rate = pricing_data[service_name]['standard_hourly_rate']
            instance_count = service_config['instances']
            service_cost += hourly_rate * instance_count * HOURS_PER_MONTH
        
        # Usage-based costs (standard rates)
        if 'usage_metrics' in service_config:
            for metric, usage in service_config['usage_metrics'].items():
                unit_cost = pricing_data[service_name]['standard_usage'][metric]
                service_cost += usage * unit_cost
        
        cost_breakdown[service_name] = round(service_cost, 2)
        total_cost += service_cost
    
    return {
        'total_monthly_cost': round(total_cost, 2),
        'service_breakdown': cost_breakdown
    }

Implementation Checklist

Credit Exclusion (MANDATORY FIRST)

• [ ] Excluded ALL credits from native tool analysis
• [ ] Calculated gross monthly cost (true infrastructure cost)
• [ ] Assessed credit expiration timeline

Analysis Steps

• [ ] Used native cost tools or CLI commands with credits excluded for recent 3 complete months
• [ ] Applied correct CLI filters for each provider (AWS: RECORD_TYPE, Azure: ChargeType, GCP: BigQuery)
• [ ] Specified exact date range in analysis
• [ ] Discovered all resources across all environments
• [ ] Researched current standard pricing rates
• [ ] Calculated costs using Python with standard rates
• [ ] Validated native vs calculated costs using gross amounts from recent months

Output Format

1. Credit Impact Analysis

• Analysis Period: [Specify actual 3 months analyzed]
• Gross Monthly Cost (without credits): $X,XXX - PRIMARY NUMBER
• Net Monthly Cost (with credits): $X,XXX
• Credits Applied: $XXX/month
• Credit Expiration Risk: Timeline assessment

2. True Infrastructure Costs

• Total Monthly Cost: Min/avg/max scenarios at standard pricing
• Service Breakdown: Cost per service without discounts
• Environment Breakdown: Cost per environment at standard rates

3. Validation & Assumptions

• Native gross cost vs calculated cost comparison
• Key assumptions and methodology
• Python scripts for reproducibility

Critical Reminders

• ALWAYS EXCLUDE CREDITS FIRST - Use specific CLI commands for each provider
• USE RECENT 3 COMPLETE MONTHS - Calculate current date, then use last 3 complete months
• Correct CLI Filters: AWS (RECORD_TYPE), Azure (ChargeType + REST API), Google Cloud (BigQuery only)
• Use native tools - Most accurate real-time data
• Standard pricing only - No promotional rates
• Python for all math - Prevent calculation errors
• Include ALL resources - Incomplete discovery causes surprises
• Document assumptions - Enable validation and updates