Module 4: Rate Optimization

Duration: 60 minutes | Level: Tactical + Strategic
WAF Alignment: CO:05 (Best Rates), CO:06 (Billing Increments)

4.1 Rate Optimization Overview

Rate optimization is the practice of reducing Azure unit costs without changing workload behavior. While usage optimization eliminates waste, rate optimization ensures you pay the lowest possible price for every resource you actually need. Microsoft offers multiple discount mechanisms — the key is selecting the right combination for each workload.

Key Principle: Combine multiple strategies for maximum savings. For example, a Windows SQL VM can benefit from Azure Hybrid Benefit (Windows OS) + AHB (SQL license) + Reserved Instance = up to 82% total savings.

📖 Azure Well-Architected Framework: CO:05 – Get the best rates

4.2 Azure Reservations vs Savings Plans — Deep Comparison

Feature Comparison

Feature	Azure Reservations	Azure Savings Plans
Maximum Savings	Up to 72%	Up to 65%
Term Options	1 year or 3 years	1 year or 3 years
Flexibility	Fixed SKU + Region	Any VM family, any region
Scope Options	Shared, Subscription, Resource Group, Management Group	Shared, Subscription, Resource Group, Management Group
Applies To	Compute, Database, Storage, Analytics, and more	Compute only (VMs, Dedicated Hosts, Container Instances, App Service, Azure Functions Premium)
Auto-apply	Yes (auto-applied to matching usage)	Yes (auto-applied to eligible compute)
Instance Size Flexibility	Yes (within same VM series/region)	Inherent (any size, any series)
Exchange Policy	Exchange for same type reservation (value must be ≥ original). Exchanges retired for compute after Oct 2024 — use Savings Plans instead	Not exchangeable
Refund/Cancellation	Self-service cancel with early termination fee. Lifetime refund limit of $50,000 USD per enrollment	Self-service cancel with early termination fee. Lifetime refund limit of $50,000 USD per enrollment
Payment Options	Upfront, Monthly, or combination	Upfront or Monthly
Best For	Stable, predictable workloads staying on same SKU/region	Variable compute across families/regions

⚠️ Important Policy Change: As of October 2024, reservation exchanges for compute services are retired. Microsoft recommends purchasing Savings Plans for flexibility going forward. Non-compute reservations (databases, storage) can still be exchanged.

📖 Changes to the Azure reservation exchange policy

When to Stack Both

You can hold both Reservations and Savings Plans simultaneously. Azure applies discounts in this priority order:

Reservations (applied first to exact-match usage)
Savings Plans (applied next to remaining eligible compute)
Pay-as-you-go (remaining usage)

This means you can buy Reservations for your most stable workloads and a Savings Plan as a "safety net" for variable compute — maximizing discount coverage.

Decision Matrix

📖 Azure Reservations documentation
📖 Azure Savings Plans documentation

4.3 Reservation Utilization Monitoring & Alerts

Low reservation utilization means you are paying for committed capacity that sits unused. Target >95% utilization across all reservations.

How to Monitor Utilization

In the Azure Portal:

Navigate to Cost Management + Billing > Reservations
Click on any reservation to see its Utilization (%) over the past 7/30 days
Azure Advisor will flag reservations below 100% utilization as a recommendation

Via Azure CLI:

# List all reservations and their utilization summary
az reservations reservation list \
  --reservation-order-id <order-id> \
  --output table

# Get reservation utilization details
az consumption reservation summary list \
  --reservation-order-id <order-id> \
  --grain daily \
  --start-date 2025-01-01 \
  --end-date 2025-01-31 \
  --output table

Via Azure Resource Graph (KQL):

// Find reservations with low utilization via Advisor
AdvisorResources
| where type == "microsoft.advisor/recommendations"
| where properties.category == "Cost"
| where properties.shortDescription.solution contains "reservation"
| project
    Recommendation = tostring(properties.shortDescription.solution),
    Impact = tostring(properties.impact),
    ResourceId = tostring(properties.resourceMetadata.resourceId)

Setting Up Utilization Alerts

Go to Cost Management + Billing > Reservations
Click Alerts in the left menu
Configure an alert rule:
- Utilization threshold: Alert when utilization drops below 80% (or 90%)
- Granularity: Daily or weekly
- Notification: Email to FinOps team or Action Group
Alternatively, create a Budget Alert scoped to reservation charges

📖 Monitor reservation utilization
📖 Set up reservation utilization alerts

4.4 Services Supporting Reservations & Reserved Capacity

Compute Reservations

Service	Savings	Notes
Virtual Machines	Up to 72%	Instance size flexibility within same series
VM Scale Sets (VMSS)	Up to 72%	Same reservation as VMs
Dedicated Hosts	Up to 60%	Entire physical server
Azure VMware Solution	Up to 60%	Dedicated VMware nodes
Azure Red Hat OpenShift	Up to 50%	Worker node compute

Database Reserved Capacity

Service	Savings	Commitment
Azure SQL Database	Up to 55%	vCore-based, 1 or 3 year
Azure SQL Managed Instance	Up to 55%	vCore-based, 1 or 3 year
Azure Cosmos DB	Up to 65%	Request Units (RU/s), 1 or 3 year
Azure Database for MySQL	Up to 55%	vCore-based, 1 or 3 year
Azure Database for PostgreSQL	Up to 55%	vCore-based, 1 or 3 year
Azure Database for MariaDB	Up to 55%	vCore-based, 1 or 3 year
Azure Cache for Redis	Up to 55%	Node-based, 1 or 3 year

Storage Reservations

Service	Savings	Commitment
Azure Blob Storage	Up to 38%	100 TB or 1 PB blocks, 1 or 3 year
Azure Data Lake Storage	Up to 38%	100 TB or 1 PB blocks, 1 or 3 year
Azure Files	Up to 36%	10 TiB or 100 TiB blocks
Azure Managed Disks	Up to 30%	P30/P40/P50 Premium SSD, 1 year

Analytics & Other Reservations

Service	Savings	Notes
Azure Synapse Analytics	Up to 60%	cDWU-based commitment
Azure Databricks	Up to 38%	DBU-based commitment
Azure Data Explorer	Up to 52%	Markup units, 1 or 3 year
Azure App Service (Premium v3)	Up to 55%	Stamp fee savings
Azure Backup	Up to 25%	Protected instances
Azure Data Factory	Up to 40%	Data flow vCore-hours
Microsoft Fabric	Up to 40%	Capacity Units, 1 year

📖 Products that support reservations
📖 Azure SQL reserved capacity
📖 Cosmos DB reserved capacity
📖 Azure Cache for Redis reservations

4.5 Azure Hybrid Benefit (AHB) — Comprehensive Guide

What Licenses Qualify?

License Type	Azure Benefit	Max Savings
Windows Server (with SA or subscription)	Free Windows OS on Azure VMs	Up to 40%
SQL Server (with SA or subscription)	Free SQL license on Azure SQL DB, SQL MI, SQL VM	Up to 55%
Linux (RHEL/SUSE subscriptions)	Discounted subscription pricing	Up to 60%
Combined (Windows + SQL Server)	Stacked benefits on same VM	Up to 82%
AKS Windows Containers	Free Windows node licensing	Up to 40%
Azure Stack HCI	Free Azure Stack HCI host licensing	Up to 40%

AHB for AKS (Windows Containers)

Azure Hybrid Benefit applies to Windows node pools in AKS. Each Windows Server core license with active Software Assurance covers up to one Windows node in AKS. This eliminates the per-vCPU Windows licensing surcharge.

How to enable AHB on an AKS cluster:

# Enable AHB when creating a new AKS cluster with Windows node pool
az aks create \
  --resource-group myRG \
  --name myAKSCluster \
  --windows-admin-username azureuser \
  --windows-admin-password $WIN_PASSWORD \
  --node-count 1 \
  --enable-ahub

# Enable AHB on an existing AKS cluster
az aks update \
  --resource-group myRG \
  --name myAKSCluster \
  --enable-ahub

# Disable AHB on an AKS cluster
az aks update \
  --resource-group myRG \
  --name myAKSCluster \
  --disable-ahub

📖 Azure Hybrid Benefit for AKS

AHB for Azure Stack HCI

Azure Stack HCI customers with Windows Server Datacenter licenses (with SA) can use AHB to avoid paying the Azure Stack HCI host fee. This benefit applies to the host-level licensing for running Azure Stack HCI clusters.

Key Points:

Requires Windows Server Datacenter with active Software Assurance or subscription licenses
Covers the Azure Stack HCI host fee (currently billed per physical core per month)
Must be enabled per-cluster through the Azure portal or Azure CLI
Can be combined with Azure Hybrid Benefit for guest Windows Server VMs running on the cluster

📖 Azure Hybrid Benefit for Azure Stack HCI

Tracking AHB Using Azure Workbooks

Azure provides a Hybrid Benefit Tracking Workbook to monitor AHB usage across your entire estate:

Navigate to Azure Portal > Azure Advisor > Workbooks
Open the Azure Hybrid Benefit workbook (or search in the Workbooks gallery)
The workbook shows:
- Total VMs with and without AHB enabled
- Potential savings from enabling AHB on remaining VMs
- License utilization vs. entitlements
- Breakdown by subscription, resource group, and VM size
You can also build a custom workbook using this Azure Resource Graph query:

// Find VMs without AHB enabled
Resources
| where type == "microsoft.compute/virtualmachines"
| where properties.storageProfile.imageReference.publisher == "MicrosoftWindowsServer"
| extend licenseType = tostring(properties.licenseType)
| summarize
    AHB_Enabled = countif(licenseType == "Windows_Server"),
    AHB_Not_Enabled = countif(licenseType != "Windows_Server")
| extend Total = AHB_Enabled + AHB_Not_Enabled
| extend Pct_Without_AHB = round(todouble(AHB_Not_Enabled) / todouble(Total) * 100, 1)

📖 Azure Hybrid Benefit Workbook

AHB Enforcement Policies

The following Azure Policy enforces Azure Hybrid Benefit for Windows VMs:

{
  "properties": {
    "displayName": "Enforce Azure Hybrid Benefit for Windows",
    "policyType": "Custom",
    "mode": "All",
    "description": "Deny Windows VMs without Azure Hybrid Benefit enabled",
    "policyRule": {
      "if": {
        "allOf": [
          {
            "field": "type",
            "in": ["Microsoft.Compute/virtualMachines",
                    "Microsoft.Compute/VirtualMachineScaleSets"]
          },
          {
            "field": "Microsoft.Compute/imagePublisher",
            "in": ["MicrosoftWindowsServer", "MicrosoftWindowsDesktop"]
          },
          {
            "field": "Microsoft.Compute/licenseType",
            "notin": ["Windows_Server", "Windows_Client"]
          }
        ]
      },
      "then": { "effect": "deny" }
    }
  }
}

AHB for SQL VMs Policy

{
  "properties": {
    "displayName": "Enforce Azure Hybrid Benefit for SQL",
    "policyType": "Custom",
    "mode": "All",
    "description": "Deny SQL VMs without Azure Hybrid Benefit enabled",
    "policyRule": {
      "if": {
        "allOf": [
          { "field": "type", "equals": "Microsoft.SqlVirtualMachine/SqlVirtualMachines" },
          { "field": "Microsoft.SqlVirtualMachine/SqlVirtualMachines/sqlImageSku",
            "in": ["Standard", "Enterprise"] },
          { "field": "Microsoft.SqlVirtualMachine/SqlVirtualMachines/sqlServerLicenseType",
            "notequals": "AHUB" }
        ]
      },
      "then": { "effect": "deny" }
    }
  }
}

Available in knowledge base: Module Rate Optimization/Policy-Enforce-AHB-Windows.json, Policy-Enforce-AHB-SQLVMs.json

Checking AHB Status via Azure CLI

# Check AHB status for all VMs in a subscription
az vm list \
  --query "[].{Name:name, RG:resourceGroup, LicenseType:licenseType, OSType:storageProfile.osDisk.osType}" \
  --output table

# Find all Windows VMs WITHOUT AHB enabled
az vm list \
  --query "[?storageProfile.imageReference.publisher=='MicrosoftWindowsServer' && licenseType!='Windows_Server'].{Name:name, RG:resourceGroup, LicenseType:licenseType}" \
  --output table

# Enable AHB on a specific VM
az vm update \
  --resource-group myRG \
  --name myVM \
  --license-type Windows_Server

# Bulk-enable AHB on all Windows VMs in a resource group
az vm list --resource-group myRG \
  --query "[?storageProfile.imageReference.publisher=='MicrosoftWindowsServer' && licenseType!='Windows_Server'].name" \
  --output tsv | \
  xargs -I {} az vm update --resource-group myRG --name {} --license-type Windows_Server

# Check AHB status for SQL VMs
az sql vm list \
  --query "[].{Name:name, RG:resourceGroup, LicenseType:sqlServerLicenseType}" \
  --output table

📖 Azure Hybrid Benefit overview
📖 Azure Hybrid Benefit for Windows Server
📖 Azure Hybrid Benefit for SQL Server

4.6 Spot VMs — Deep Dive

Spot VM Overview

Azure Spot VMs let you access unused Azure compute capacity at discounts of up to 90% compared to pay-as-you-go pricing. In return, Azure can evict your VM when it needs the capacity back.

Eviction Types

Eviction Type	Behavior	When to Use
Capacity-Only	VM is evicted only when Azure needs the capacity back. Price does not trigger eviction.	When you want to keep running as long as capacity exists, regardless of price fluctuations
Price-or-Capacity	VM is evicted if Azure needs capacity back OR if the Spot price exceeds your maximum price. You set a max price you are willing to pay.	When you want cost control — will not pay above a certain threshold

Eviction Policies

Policy	What Happens on Eviction	Use Case
Deallocate	VM is stopped and deallocated. Disks are retained. VM can be restarted later (if capacity available). You continue to pay for disk storage.	Batch jobs where you want to resume later
Delete	VM and its disks are fully deleted. Nothing is retained.	Ephemeral/stateless workloads, CI/CD runners

Configuration Examples

# Create a Spot VM with Price-or-Capacity eviction, max price $0.05/hr
az vm create \
  --resource-group myRG \
  --name mySpotVM \
  --image Ubuntu2204 \
  --size Standard_D4s_v5 \
  --priority Spot \
  --eviction-policy Deallocate \
  --max-price 0.05

# Create a Spot VM with Capacity-Only eviction (max price = -1 means up to PAYG rate)
az vm create \
  --resource-group myRG \
  --name mySpotVM2 \
  --image Ubuntu2204 \
  --size Standard_D4s_v5 \
  --priority Spot \
  --eviction-policy Delete \
  --max-price -1

# Check current Spot pricing history for a VM size
az vm list-skus \
  --location eastus \
  --resource-type virtualMachines \
  --size Standard_D4s_v5 \
  --output table

Spot VM Best Practices

#	Practice	Details
1	Use Spot Priority Mix for VMSS	Configure 80% Spot / 20% regular in VMSS for cost-effective reliability
2	Implement checkpointing	Save progress periodically for long-running batch jobs
3	Multi-region deployment	Spread Spot VMs across regions to reduce eviction likelihood
4	Handle eviction signals	Azure sends a 30-second eviction notice via Metadata Service — hook into this programmatically
5	Combine with on-demand	Use Spot for burst capacity, on-demand for baseline
6	AKS Spot node pools	Add Spot node pools for non-critical and batch workloads on Kubernetes
7	Avoid single-SKU dependence	Use multiple VM sizes in VMSS to increase Spot allocation success

Handling Eviction Notifications

# Query the Azure Instance Metadata Service for eviction notice (run inside the VM)
curl -H "Metadata:true" \
  "http://169.254.169.254/metadata/scheduledevents?api-version=2020-07-01"

📖 Spot VMs overview
📖 Spot VMs for VMSS
📖 AKS Spot node pools

4.7 On-Demand Capacity Reservations

On-Demand Capacity Reservations guarantee compute capacity in a specific Azure region without any term commitment. Unlike Reserved Instances (which discount price but do not guarantee capacity), Capacity Reservations guarantee capacity but do not provide a price discount.

Feature	On-Demand Capacity Reservation	Reserved Instances
Capacity Guarantee	Yes (guaranteed in-region)	No (best-effort)
Price Discount	No (full pay-as-you-go)	Yes (up to 72%)
Term	None (pay hourly)	1 or 3 years
Use Case	DR readiness, product launches, critical capacity needs	Cost savings for stable workloads

Best Practice: Combine On-Demand Capacity Reservations with Reserved Instances or Savings Plans. The RI/SP provides the price discount; the Capacity Reservation provides the capacity guarantee. Both can be applied to the same VM simultaneously.

# Create a capacity reservation group
az capacity reservation group create \
  --resource-group myRG \
  --name myCapacityGroup \
  --location eastus

# Create a capacity reservation for 5x Standard_D4s_v5
az capacity reservation create \
  --resource-group myRG \
  --capacity-reservation-group myCapacityGroup \
  --name myReservation \
  --sku Standard_D4s_v5 \
  --capacity 5 \
  --location eastus

📖 On-Demand Capacity Reservations

4.8 Software Plans (SUSE & Red Hat)

Software Plans provide discounted pricing for Linux subscription software bundled with Azure VMs. Instead of paying hourly PAYG rates for RHEL or SUSE subscriptions, you commit to a 1- or 3-year plan.

Software Plan	Savings	Covers
SUSE Linux Enterprise Server	Up to 67% vs PAYG	SLES subscription (Standard or HPC)
Red Hat Enterprise Linux	Up to 50% vs PAYG	RHEL subscription
SUSE Linux Enterprise for SAP	Up to 67% vs PAYG	SLES for SAP, includes priority support
Red Hat for SAP	Up to 44% vs PAYG	RHEL for SAP with HA and Update Services

Key Characteristics:

Purchased separately from VM reservations
Stack with VM Reserved Instances for combined savings (compute + software)
Scope can be shared across subscription or resource group
No exchange or refund after purchase

# View available software plans via Azure CLI
az reservations catalog show \
  --subscription-id <sub-id> \
  --reserved-resource-type SuseLinux \
  --location eastus

az reservations catalog show \
  --subscription-id <sub-id> \
  --reserved-resource-type RedHat \
  --location eastus

📖 SUSE software plans
📖 Red Hat software plans

4.9 Dev/Test Pricing — Complete Guide

Dev/Test subscriptions provide discounted (or free) pricing for non-production workloads. There is no SLA for Dev/Test subscriptions.

Complete List of Discounted Services

Service	Dev/Test Discount
Windows VMs	Pay Linux rates only (no Windows license charge)
Azure SQL Database	Up to 55% discount (no SQL license charge for vCore)
Azure SQL Managed Instance	Up to 55% discount (no SQL license charge for vCore)
Logic Apps	Up to 50% discount on enterprise connectors
App Service (Basic, Standard, Premium)	Reduced rates (varies by tier)
Azure Cloud Services	Discounted compute pricing
HDInsight	No Windows node surcharge
Azure DevTest Labs	Auto-shutdown, quota management, artifact repos (free orchestration layer)
BizTalk Services	Discounted processing
Virtual Network (VPN Gateway)	Lower gateway rates
Azure API Management	Reduced pricing on Dev/Test tiers

Subscription Types

Subscription Offer	Who Can Use	How to Get
Enterprise Dev/Test	EA customers only	Create under EA enrollment in Azure EA Portal
Pay-As-You-Go Dev/Test	Visual Studio subscribers	Create in Azure portal with VS subscription credentials
MSDN Platforms	MSDN subscribers	Monthly Azure credits + Dev/Test pricing

How to Enable

Ensure the user has an active Visual Studio subscription (for PAYG Dev/Test) or an EA enrollment (for Enterprise Dev/Test)
Create a new subscription under the appropriate offer type
Place the subscription under the correct Management Group (e.g., "Non-Production")
Deploy dev, test, staging, and sandbox workloads there
Never deploy production workloads — no SLA, compliance, or support guarantees

📖 Azure Dev/Test pricing
📖 Enterprise Dev/Test offer

4.10 Commitment-Based Pricing for Azure OpenAI

Azure OpenAI offers Provisioned Throughput Units (PTUs) with commitment-based pricing for organizations with predictable, high-volume AI workloads.

Pricing Model	How It Works	Best For
Pay-as-you-go (Token-based)	Pay per 1K tokens consumed. No commitment.	Variable, low-to-moderate usage
Provisioned Throughput (PTU) — Monthly	Reserve dedicated throughput capacity. Monthly commitment.	Predictable, high-volume production workloads
Provisioned Throughput (PTU) — Yearly	Same as monthly but with annual commitment. Deeper discount.	Stable, enterprise-scale AI deployments

Key Details:

PTU commitment provides guaranteed throughput (tokens-per-minute) with latency SLAs
Monthly PTU reservations offer discounts over PAYG; yearly commitments yield greater discounts
You select a model (e.g., GPT-4, GPT-4o) and region — pricing varies by model
Minimum commitment is typically 50–100 PTUs depending on the model
Unused PTUs are still billed — right-size your commitment to actual usage

When to Use PTU Pricing:

Production applications serving >1M tokens/minute
Applications requiring consistent low-latency responses
Batch processing with guaranteed throughput windows
Cost predictability for AI spend in budgeting and forecasting

📖 Azure OpenAI Provisioned Throughput
📖 Azure OpenAI pricing

4.11 B-Series Burstable VMs

B-Series VMs are economical burstable VMs designed for workloads that do not need continuous full CPU performance. They use a CPU credit model — the VM earns credits during low-usage periods and spends them during bursts.

How the Credit Model Works

Key Characteristics

Feature	Detail
Pricing	20–60% cheaper than equivalent D-series
Baseline CPU	Typically 5–40% of full vCPU performance (varies by SKU)
Credit Earning	Continuously earned when running below baseline
Credit Cap	Each SKU has a maximum credit balance
Burst Duration	Depends on credits banked and burst intensity
Persistent Credits	Credits persist while VM is running; lost on deallocation

When to Use B-Series

Good Fit	Poor Fit
Dev/test machines	High-CPU production workloads
Low-traffic web servers	Database servers under constant load
Build agents (idle between builds)	Machine learning training
Small databases with occasional queries	Video encoding
Microservices with sporadic traffic	Real-time analytics
Jump boxes / bastion hosts	Sustained high-throughput processing

# List available B-series SKUs in a region
az vm list-skus --location eastus \
  --resource-type virtualMachines \
  --query "[?starts_with(name, 'Standard_B')].[name, capabilities[?name=='vCPUs'].value | [0], capabilities[?name=='MemoryGB'].value | [0]]" \
  --output table

# Create a B-series VM
az vm create \
  --resource-group DevTestRG \
  --name dev-web-01 \
  --image Ubuntu2204 \
  --size Standard_B2s \
  --admin-username azureuser \
  --generate-ssh-keys

📖 B-series burstable VM sizes

4.12 Rate Optimization Implementation Roadmap

Detailed Timeline

Phase	Timeline	Actions	Expected Savings
1. Assessment	Week 1–2	Inventory all compute, DB, and storage. Review Azure Advisor cost recommendations. Analyze 30/60-day usage patterns. Identify AHB-eligible VMs.	Baseline established
2. Quick Wins	Week 3–4	Enable AHB on all eligible Windows/SQL VMs and AKS clusters. Deploy deny policies. Move non-production workloads to Dev/Test subscriptions. Switch dev VMs to B-series.	15–30% on eligible VMs
3. Commitments	Week 5–8	Purchase Reserved Instances for top 10 stable VM SKUs. Purchase Savings Plans for variable compute. Buy reserved capacity for databases (SQL, Cosmos DB). Purchase software plans (SUSE/RHEL).	30–65% on committed
4. Spot & Tune	Week 9–12	Deploy Spot node pools in AKS. Configure Spot Priority Mix for VMSS. Monitor reservation utilization (target >95%). Adjust scope/size as needed. Evaluate Azure OpenAI PTU.	Additional 10–30% on batch/burst
5. Ongoing	Monthly	Review reservation utilization alerts. Assess new Advisor recommendations. Right-size commitments quarterly. Track AHB compliance via workbook.	Maintain optimized state

4.13 Rate Optimization Checklist

#	Action	Priority
1	Review Azure Advisor reservation and savings plan recommendations	High
2	Analyze 30/60-day usage patterns for commitment decisions	High
3	Enable Azure Hybrid Benefit on all eligible Windows VMs	High
4	Enable Azure Hybrid Benefit on all eligible SQL VMs	High
5	Enable Azure Hybrid Benefit on AKS Windows node pools	High
6	Deploy AHB enforcement policies (deny without AHB)	High
7	Set up AHB tracking workbook for compliance visibility	Medium
8	Move dev/test workloads to Dev/Test subscriptions	High
9	Switch dev/test VMs to B-series burstable SKUs	Medium
10	Purchase Reserved Instances for stable compute workloads	High
11	Purchase Savings Plans for variable compute spend	High
12	Purchase reserved capacity for stable databases (SQL, Cosmos DB, MySQL, PostgreSQL)	Medium
13	Purchase SUSE/RHEL Software Plans where applicable	Medium
14	Evaluate Spot VMs for batch, CI/CD, and non-critical workloads	Medium
15	Configure Spot Priority Mix in VMSS (80/20 Spot/On-demand)	Medium
16	Set up On-Demand Capacity Reservations for DR / critical launches	Low
17	Monitor reservation utilization weekly (target >95%)	High
18	Set up reservation utilization alerts (threshold below 80%)	High
19	Evaluate Azure OpenAI PTU commitment for high-volume AI workloads	Low
20	Review and re-optimize commitments quarterly	High

References

Previous Module: Module 3 — Financial Controls & Budgets
Next Module: Module 5 — Usage Optimization & Waste Reduction
Back to Overview: README — Cost Optimization

4.1 Rate Optimization Overview​

4.2 Azure Reservations vs Savings Plans — Deep Comparison​

Feature Comparison​

When to Stack Both​

Decision Matrix​

4.3 Reservation Utilization Monitoring & Alerts​

How to Monitor Utilization​

Setting Up Utilization Alerts​

4.4 Services Supporting Reservations & Reserved Capacity​

Compute Reservations​

Database Reserved Capacity​

Storage Reservations​

Analytics & Other Reservations​

4.5 Azure Hybrid Benefit (AHB) — Comprehensive Guide​

What Licenses Qualify?​

AHB for AKS (Windows Containers)​

AHB for Azure Stack HCI​

Tracking AHB Using Azure Workbooks​

AHB Enforcement Policies​

AHB for SQL VMs Policy​

Checking AHB Status via Azure CLI​

4.6 Spot VMs — Deep Dive​

Spot VM Overview​

Eviction Types​

Eviction Policies​

Configuration Examples​

Spot VM Best Practices​

Handling Eviction Notifications​

4.7 On-Demand Capacity Reservations​

4.8 Software Plans (SUSE & Red Hat)​

4.9 Dev/Test Pricing — Complete Guide​

Complete List of Discounted Services​

Subscription Types​

How to Enable​

4.10 Commitment-Based Pricing for Azure OpenAI​

4.11 B-Series Burstable VMs​

How the Credit Model Works​

Key Characteristics​

When to Use B-Series​

4.12 Rate Optimization Implementation Roadmap​

Detailed Timeline​

4.13 Rate Optimization Checklist​

References​