Azure Cost Management

Provision Resource → Meters Track Usage → Generate Usage Record → Calculate Bill

VM Cost = (Compute Hours × Compute Rate) +
          (Storage GB × Storage Rate) +
          (Network Egress GB × Network Rate) +
          (License Costs if applicable)

Standard_D4s_v3 (4 vCPUs, 16 GB RAM) - East US
├── Compute: $0.192/hour
├── Storage: Premium SSD 128GB: $19.71/month
├── Network: Outbound data transfer: $0.087/GB (first 5GB free)
└── License: Windows Server: +$0.096/hour
    Total: ~$207/month (running 24/7)

Same VM with Reserved Instance (1-year):
└── Total: ~$124/month (40% savings)

Same VM with Reserved Instance (3-year):
└── Total: ~$89/month (57% savings)

Pay-as-you-go Pricing:
├── Azure SQL Database (4 vCores, Gen5)
├── Cost: $1.46/hour
└── Monthly Cost (24/7): $1,051

Reserved Capacity (1-year commitment):
├── Discount: 33% off pay-as-you-go
└── Monthly Cost: $704 (saving $347/month)

Reserved Capacity (3-year commitment):
├── Discount: 62% off pay-as-you-go
└── Monthly Cost: $399 (saving $652/month)

{
  "profiles": [
    {
      "name": "Business hours",
      "capacity": {
        "minimum": 2,
        "maximum": 10,
        "default": 2
      },
      "rules": [
        {
          "metricTrigger": {
            "metricName": "CpuPercentage",
            "operator": "GreaterThan",
            "threshold": 75,
            "timeWindow": "PT5M"
          },
          "scaleAction": {
            "direction": "Increase",
            "value": 2
          }
        },
        {
          "metricTrigger": {
            "metricName": "CpuPercentage",
            "operator": "LessThan",
            "threshold": 25,
            "timeWindow": "PT5M"
          },
          "scaleAction": {
            "direction": "Decrease",
            "value": 1
          }
        }
      ],
      "recurrence": {
        "frequency": "Week",
        "schedule": {
          "timeZone": "Pacific Standard Time",
          "days": ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday"],
          "hours": [8],
          "minutes": [0]
        }
      }
    },
    {
      "name": "Off hours",
      "capacity": {
        "minimum": 1,
        "maximum": 2,
        "default": 1
      }
    }
  ]
}

Original VM Deployment Creates:
├── Virtual Machine
├── OS Disk (Premium SSD 128GB): $19.71/month
├── Network Interface: Minimal cost
├── Public IP Address: $3.65/month
├── Network Security Group: Free
└── Virtual Network: Free (within limits)

After VM Deletion (if not properly cleaned):
├── OS Disk: Still charged $19.71/month ❌
├── Network Interface: Still exists ❌
├── Public IP Address: Still charged $3.65/month ❌
└── Orphaned snapshots: Additional charges ❌

Total wasted cost: $23.36/month per forgotten VM

az vm delete \
  --resource-group MyResourceGroup \
  --name MyVM \
  --yes

az disk delete --ids $(az disk list --resource-group MyResourceGroup --query "[?contains(name,'MyVM')].id" -o tsv)
az network nic delete --ids $(az network nic list --resource-group MyResourceGroup --query "[?contains(name,'MyVM')].id" -o tsv)
az network public-ip delete --ids $(az network public-ip list --resource-group MyResourceGroup --query "[?contains(name,'MyVM')].id" -o tsv)

Decision Framework:
├── Data Residency Requirements
│   └── Must store data in specific geography for compliance
├── User Location
│   └── Choose region closest to end users (latency < 50ms optimal)
├── Service Availability
│   └── Not all services available in all regions
├── Disaster Recovery
│   └── Pair primary region with appropriate secondary
└── Cost Optimization
    └── Select cheapest region meeting above requirements

Application: E-commerce platform
Primary Users: United States
Requirements: 
├── Low latency for users
├── Data sovereignty: US data must stay in US
├── High availability required
└── Cost optimization important

Recommended Architecture:
├── Primary Region: East US (lowest US cost)
├── Secondary Region: West US 2 (DR, same cost)
├── CDN: Azure Front Door (global)
└── Database Replicas: Read replicas in East US only
    (avoid cross-region replication costs)

Cost Savings vs Suboptimal Selection:
└── Avoided Brazil South: Saved 47% on compute
└── Avoided West Europe: Saved 14% on compute
└── Single region database: Saved ~$200/month on geo-replication

Inbound Data Transfer (Ingress):
└── Free for all scenarios ✅

Outbound Data Transfer (Egress) - Monthly Tiers:
├── First 5 GB: Free
├── 5 GB - 10 TB: $0.087/GB (Zone 1)
├── 10 TB - 50 TB: $0.083/GB (Zone 1)
├── 50 TB - 150 TB: $0.070/GB (Zone 1)
└── 150 TB - 500 TB: $0.050/GB (Zone 1)

Cross-Zone Pricing (more expensive):
├── Zone 1 to Zone 2: $0.120/GB
└── Zone 1 to Zone 3: $0.150/GB

Data Transfer Within Azure:
├── Same Region: Free ✅
├── Same Zone (different regions): $0.02/GB
└── Different Zones: Regular egress rates

Original Architecture (High Cost):
├── Web servers: East US
├── Database: West Europe (geo-replicated)
├── Object storage: Australia East
├── CDN: Not used
└── Monthly data egress: 100 TB

Cost Breakdown:
├── Database cross-region replication: 50 TB/month × $0.02 = $1,000
├── Application to database traffic: 30 TB/month × $0.02 = $600
├── Direct video streaming: 100 TB × $0.083 = $8,300
└── Total Network Cost: $9,900/month

Optimized Architecture (Low Cost):
├── Web servers: East US
├── Database: East US (local only)
├── Object storage: East US
├── CDN: Azure Front Door with caching
└── Monthly data egress: 100 TB (via CDN, discounted)

Cost Breakdown:
├── Database cross-region replication: Eliminated ✅
├── Application to database traffic: Free (same region) ✅
├── CDN egress: 100 TB × $0.067 (CDN discount) = $6,700
└── Total Network Cost: $6,700/month

Monthly Savings: $3,200 (32% reduction)

Includes:
├── $200 credit for first 30 days
├── 12 months of popular services free:
│   ├── 750 hours B1s Linux VM
│   ├── 750 hours B1s Windows VM
│   ├── 5 GB blob storage
│   ├── 250 GB SQL Database
│   └── 15 GB outbound data transfer
└── 25+ always-free services:
    ├── 10 web, mobile, or API apps
    ├── 1 million Azure Functions executions
    └── 20 compute hours Azure Container Instances

Best for: Learning, prototyping, small projects

Characteristics:
├── No upfront commitment
├── Pay only for what you use
├── Full pricing (no discounts)
└── Maximum flexibility

Best for: Small businesses, variable workloads, short-term projects

Benefits:
├── Volume discounts (10-40% typically)
├── Azure Monetary Commitment
├── Flexible payment options
├── Centralized billing
├── Reserved Instance discounts stackable
└── Access to Azure Dev/Test pricing

Best for: Large organizations, predictable annual spend >$100K

Discounts:
├── Windows VMs: Save ~40% (no Windows license fee)
├── SQL Server VMs: Save ~55% (no SQL license fee)
├── Azure SQL Database: Save ~55%
└── Other compute services: Various discounts

Requirements:
├── Must be used for non-production workloads only
├── Requires Visual Studio subscription
└── Subject to audit

Best for: Development, testing, staging environments

public class SubscriptionCostAnalyzer
{
    public decimal CalculatePotentialSavings(
        List<Resource> resources, 
        SubscriptionType currentType)
    {
        decimal currentCost = 0;
        decimal optimizedCost = 0;
        
        foreach (var resource in resources)
        {
            currentCost += GetResourceCost(resource, currentType);
            
            // Determine optimal subscription type for this resource
            var optimalType = GetOptimalSubscriptionType(resource);
            optimizedCost += GetResourceCost(resource, optimalType);
        }
        
        return currentCost - optimizedCost;
    }
    
    private SubscriptionType GetOptimalSubscriptionType(Resource resource)
    {
        if (resource.IsProduction && resource.RunsConstantly)
        {
            return SubscriptionType.ReservedInstance;
        }
        else if (resource.IsDevelopment || resource.IsTest)
        {
            return SubscriptionType.DevTest;
        }
        else if (resource.IsSpiky || resource.IsShortTerm)
        {
            return SubscriptionType.PayAsYouGo;
        }
        
        return SubscriptionType.PayAsYouGo;
    }
}

Marketplace Solution Cost Breakdown:
├── Azure Infrastructure:
│   ├── VM Compute: Standard_D4s_v3: $0.192/hour
│   ├── Storage: Premium SSD 128GB: $19.71/month
│   ├── Public IP: $3.65/month
│   └── Data Transfer: Variable based on usage
│
└── Vendor Software License:
    ├── FortiGate License: $0.35/hour
    └── Support & Updates: Included
    
Total Hourly Cost: $0.192 + $0.35 = $0.542/hour
Total Monthly Cost: ~$396/month

Compare to:
├── Building your own firewall: Lower cost, higher effort
├── Managed Azure Firewall: $1.25/hour (~$912/month)
└── Marketplace provides middle ground

Standard Marketplace SQL Server VM:
├── VM Compute: $0.192/hour
├── SQL Server License: $0.967/hour
└── Total: $1.159/hour = $846/month

With BYOL (if you already own SQL licenses):
├── VM Compute: $0.192/hour
├── SQL Server License: $0/hour (using existing license)
└── Total: $0.192/hour = $140/month

Savings: $706/month (83% reduction on this component)

Application Requirements:
├── Frontend: Static website
├── API Layer: .NET Core application
├── Database: SQL Server
├── Storage: Images and documents
└── Expected Traffic: 100K users/month

Pricing Calculator Inputs:

1. Azure App Service:
   ├── Tier: Premium V3 (P1v3)
   ├── Instances: 2 (for high availability)
   ├── Compute Hours: 730/month
   └── Estimated Cost: $200/month

2. Azure SQL Database:
   ├── Model: vCore-based (General Purpose)
   ├── Hardware: Gen5, 4 vCores
   ├── Storage: 256 GB
   ├── Backup: 7-day retention
   └── Estimated Cost: $730/month
   
3. Azure Blob Storage:
   ├── Performance: Standard
   ├── Redundancy: GRS
   ├── Data Stored: 500 GB
   ├── Monthly Transactions: 1M
   └── Estimated Cost: $25/month
   
4. Azure CDN:
   ├── Data Transfer: 1 TB/month
   └── Estimated Cost: $82/month
   
5. Application Insights:
   ├── Data Ingestion: 5 GB/month
   └── Estimated Cost: $11/month

Total Monthly Cost: $1,048/month
Annual Cost: $12,576/year

With Reserved Instances (3-year):
├── App Service: Save 38% = $124/month saved
├── SQL Database: Save 62% = $452/month saved
└── Total Savings: $576/month or $6,912/year

New Annual Cost: $5,664/year (55% savings)

Best Practices:
├── Start with production estimates, then add dev/test
├── Include disaster recovery costs
├── Factor in data egress for accurate network costs
├── Add 20% buffer for unexpected usage
├── Compare reserved vs pay-as-you-go scenarios
├── Export estimate for stakeholder review
└── Update estimate quarterly as requirements change

Common Mistakes to Avoid:
├── Forgetting to include backups and snapshots
├── Underestimating network egress
├── Not accounting for development environments
├── Ignoring managed service overhead costs
└── Assuming linear scaling of costs

Hardware:
├── 20 physical servers (refreshed every 5 years)
├── Initial Hardware Cost: $500,000
├── Storage: 100 TB SAN: $150,000
├── Network Equipment: $50,000
└── Hardware Annual Depreciation: $140,000

Facility Costs:
├── Data Center Space: $24,000/year
├── Power & Cooling: $48,000/year
└── Physical Security: $12,000/year

IT Labor:
├── 3 Infrastructure Engineers: $300,000/year
├── 1 Database Administrator: $120,000/year
└── 0.5 Network Administrator: $60,000/year

Software Licenses:
├── Windows Server: $40,000/year
├── SQL Server: $80,000/year
├── Virtualization Platform: $30,000/year
└── Monitoring & Backup Tools: $25,000/year

Total Annual On-Premises Cost: $839,000

Over 5 Years: $4,195,000

Compute (IaaS Equivalent):
├── 20 Standard_D4s_v3 VMs
├── Pay-as-you-go: $2,803/month × 12 = $33,636/year
├── 3-year Reserved: $1,780/month × 12 = $21,360/year

Migrated to PaaS:
├── 10 App Service Plans (P2v3): $292/month × 12 = $3,504/year
├── 5 Azure SQL Databases: $365/month × 12 = $4,380/year
├── 5 Function Apps: $14/month × 12 = $168/year

Storage:
├── 100 TB Premium Storage: $4,906/month × 12 = $58,872/year
├── Backup & Archive: $800/month × 12 = $9,600/year

Networking:
├── VPN Gateway: $263/month × 12 = $3,156/year
├── Load Balancer: $18/month × 12 = $216/year
├── Data Egress (5 TB/month): $435/month × 12 = $5,220/year

Management:
├── Azure Monitor & Log Analytics: $200/month × 12 = $2,400/year
├── Azure Backup: Included in storage costs

Licenses (Hybrid Benefit):
├── Windows Server: $0 (using Azure Hybrid Benefit)
├── SQL Server: $0 (using Azure Hybrid Benefit)

Labor Reduction:
├── 2 Infrastructure Engineers: $200,000/year (1 engineer eliminated)
├── 0.5 Database Administrator: $60,000/year (50% reduction)
├── 0.25 Network Administrator: $30,000/year (50% reduction)
└── New Azure Cloud Engineer: $140,000/year

Total Annual Azure Cost (PaaS): $318,876
Total Annual Labor Cost: $430,000

Total Annual Cost: $748,876

Over 5 Years: $3,744,380

Total 5-Year Savings: $450,620 (11% reduction)

Additional Benefits Not Quantified:
├── Improved disaster recovery capabilities
├── Better scalability for growth
├── Reduced time to market for new features
├── Enhanced security and compliance
├── No hardware refresh cycles
└── Global reach and lower latency for distributed users

Cost Analysis Dimensions:
├── Resource
├── Resource Group
├── Resource Type
├── Location (Region)
├── Service Name
├── Subscription
├── Tag
├── Publisher Type (First-party vs Marketplace)
└── Reservation (Reserved vs On-demand)

Time Periods:
├── Daily
├── Monthly
├── Custom date range
└── Accumulated (running total)

Grouping and Filtering:
├── Group by resource type to see which services cost most
├── Filter by tag to see project-specific costs
├── Compare time periods (month-over-month)
└── Export to Excel for offline analysis

Query 1: Most Expensive Resources
└── Group by: Resource
└── Time: Last 30 days
└── Sort by: Cost descending
└── Result: Identifies top 10 cost drivers

Query 2: Cost Trend by Department
└── Group by: Tag (Department)
└── Time: Last 6 months
└── Chart: Line graph
└── Result: Shows departmental spending trends

Query 3: Unused Resources
└── Filter: Resource type = "Virtual Machine"
└── Metric: CPU utilization < 5%
└── Time: Last 7 days
└── Result: Candidate VMs for deletion or downsizing

Query 4: Storage Growth
└── Group by: Resource type = "Storage Account"
└── Metric: Capacity used
└── Time: Last 12 months
└── Result: Projects future storage needs

{
  "budgetName": "Engineering-Team-Q1-2026",
  "amount": 50000,
  "timeGrain": "Monthly",
  "timePeriod": {
    "startDate": "2026-01-01",
    "endDate": "2026-03-31"
  },
  "category": "Cost",
  "filter": {
    "tags": {
      "Department": "Engineering",
      "Environment": "Production"
    }
  },
  "notifications": {
    "actual_80_percent": {
      "enabled": true,
      "operator": "GreaterThan",
      "threshold": 80,
      "contactEmails": [
        "engineering-leads@company.com",
        "finance@company.com"
      ],
      "contactRoles": [
        "Owner",
        "Contributor"
      ],
      "thresholdType": "Actual"
    },
    "actual_100_percent": {
      "enabled": true,
      "operator": "GreaterThan",
      "threshold": 100,
      "contactEmails": [
        "engineering-leads@company.com",
        "cto@company.com",
        "cfo@company.com"
      ],
      "contactGroups": [
        "/subscriptions/.../resourceGroups/.../providers/microsoft.insights/actionGroups/CriticalAlerts"
      ],
      "thresholdType": "Actual"
    },
    "forecasted_110_percent": {
      "enabled": true,
      "operator": "GreaterThan",
      "threshold":