Cost Management

First PublishedFeb 17, 2026ByAtif Alam

Cloud costs grow quickly without visibility and discipline. Azure Cost Management + Billing provides the tools to understand, allocate, and optimize spending across subscriptions and resource groups.

Cost Management Overview

Tool	What It Does	AWS Equivalent
Cost Analysis	Explore costs with charts and filters	Cost Explorer
Budgets	Set spending limits with alerts	AWS Budgets
Advisor (cost)	Right-sizing and idle resource recommendations	Trusted Advisor
Pricing Calculator	Estimate costs before deploying	AWS Pricing Calculator
Azure Reservations	Pre-pay for 1/3 year discounts	Savings Plans / Reserved Instances

Cost Analysis

Cost Analysis is the primary tool for exploring your Azure spending:

Filter by:
  ├── Subscription
  ├── Resource group
  ├── Service name (e.g. Virtual Machines, Storage)
  ├── Location
  ├── Tag (e.g. Environment=production)
  └── Time range

View as:
  ├── Accumulated cost (running total)
  ├── Daily cost (bar chart)
  ├── Cost by service
  ├── Cost by resource group
  └── Cost by tag

Common Views

View	Question It Answers
Cost by service	Which Azure services cost the most?
Cost by resource group	Which teams or applications spend the most?
Cost by tag	How much does each environment (prod/dev) cost?
Daily cost	Is spending trending up or down?
Forecast	What will this month’s bill be?

Azure CLI Cost Query

# View cost for the current month by resource group
az costmanagement query \
  --type ActualCost \
  --scope "subscriptions/<subscription-id>" \
  --timeframe MonthToDate \
  --dataset-grouping name="ResourceGroup" type="Dimension"

Budgets

Budgets alert you (and optionally take action) when spending reaches thresholds:

# Create a monthly budget of $5,000
az consumption budget create \
  --budget-name "monthly-limit" \
  --amount 5000 \
  --time-grain Monthly \
  --start-date 2026-02-01 \
  --end-date 2027-01-31 \
  --resource-group myapp-prod-rg

# Alert thresholds are configured in the portal or ARM/Bicep

Budget Alert Thresholds

Threshold	Typical Action
50%	Informational email to the team
80%	Warning email to team leads
100%	Alert to finance and engineering managers
120% (forecast)	Trigger automation (e.g. stop non-critical VMs)

Budget Actions

Budgets can trigger action groups that:

• Send email/SMS notifications.
• Call a webhook.
• Trigger an Azure Function or Logic App (e.g. to shut down dev VMs).

Tagging for Cost Attribution

Tags are the foundation of cost management — without them, you can’t attribute costs to teams, applications, or environments.

Recommended Tags

Tag	Purpose	Example Values
Environment	Separate prod vs dev costs	production, staging, development
Team	Owning team	platform, backend, data
CostCenter	Finance cost center	CC-1234, engineering
Application	Application name	web-portal, api-service
ManagedBy	IaC tool	terraform, bicep, manual

Enforcing Tags

Use Azure Policy to enforce required tags — deny resource creation if tags are missing.

Cost Analysis by Tag

Once tags are applied, filter Cost Analysis by tag:

Cost by tag: Environment
  production    $12,500  (72%)
  staging        $2,800  (16%)
  development    $2,100  (12%)
  ──────────────────────
  Total         $17,400

Azure Advisor (Cost Recommendations)

Azure Advisor automatically detects optimization opportunities:

Recommendation	What It Finds
Right-size VMs	VMs with <5% average CPU — downsize or shut down
Shut down unused VMs	VMs running but with no network/disk activity
Delete unattached disks	Managed disks not attached to any VM
Use Reserved Instances	Steady-state VMs that would save with 1/3-year commit
Switch to cheaper storage tiers	Blob data that hasn’t been accessed (move to Cool/Archive)
Delete unused public IPs	Static public IPs not associated with any resource

# View cost recommendations
az advisor recommendation list --category Cost

Azure Reservations and Savings Plans

Reservations

Pre-pay for a specific resource type and size for 1 or 3 years:

Resource	1-Year Discount	3-Year Discount
VMs	~30–40%	~55–65%
SQL Database	~30%	~55%
Cosmos DB (RUs)	~20%	~40%
App Service (Isolated)	~35%	~55%

Reservations apply to a specific VM size and region. They can be scoped to a subscription, resource group, or shared across subscriptions.

Savings Plans

More flexible than reservations — commit to a dollar amount per hour across any VM size/region:

Savings Plan: $10/hour compute commitment (3-year)
  ├── Covers any VM size in any region
  ├── Covers App Service Premium
  └── Covers Azure Functions Premium
  Discount: ~50–65% (varies by resource)

Reservations vs Savings Plans

	Reservations	Savings Plans
Scope	Specific size + region	Any size, any region
Flexibility	Low (exchange/refund available)	High (automatic)
Discount	Slightly higher	Slightly lower
Best for	Stable, predictable workloads	Variable or multi-region workloads

Spot VMs

Use Azure Spot VMs for fault-tolerant workloads at up to 90% discount:

# Create a Spot VM
az vm create \
  --resource-group myapp-rg \
  --name batch-worker \
  --image Ubuntu2204 \
  --size Standard_D4s_v5 \
  --priority Spot \
  --eviction-policy Deallocate \
  --max-price 0.10

Spot VMs can be evicted when Azure needs the capacity. Use them for batch processing, CI/CD, dev/test, and stateless workloads.

Common Cost Traps

Trap	Prevention
Forgotten dev/test VMs	Auto-shutdown schedule, budget alerts, Environment tag
Oversized VMs	Start small, use Advisor right-sizing recommendations
Premium storage for dev	Use Standard SSD for non-production
Unattached disks	Periodic cleanup, Advisor alerts
Idle Application Gateways	~$175/month even with no traffic — delete when not needed
Log Analytics over-ingestion	Set daily caps, archive old data, filter noisy logs
Public IPs	Static public IPs cost ~$3.65/month each — delete unused ones
ExpressRoute / VPN Gateways	Fixed monthly cost even when idle — right-size or remove

Cost Optimization Strategies

Quick Wins

1. Auto-shutdown dev/test VMs at end of day.
2. Delete unattached disks and unused public IPs.
3. Move cold blobs to Cool or Archive tier.
4. Right-size VMs based on Advisor recommendations.
5. Use B-series VMs for bursty, low-utilization workloads.

Medium-Term

1. Buy Reservations for stable production VMs and databases.
2. Use Spot VMs for batch processing and CI/CD.
3. Azure SQL Serverless for databases with intermittent usage (auto-pause).
4. Cosmos DB autoscale instead of fixed throughput.
5. Azure Functions Consumption plan for event-driven workloads.

Organizational

1. Enforce tags via Azure Policy for cost attribution.
2. Set budgets per resource group and subscription.
3. Separate subscriptions for prod, dev, and sandbox.
4. Monthly cost reviews — review Cost Analysis with team leads.
5. Use management groups for consolidated billing visibility.

Multi-Subscription Cost Management

For organizations with multiple subscriptions:

Management Group: Company
  ├── Production MG
  │     ├── Sub: prod-web (Budget: $10,000)
  │     └── Sub: prod-data (Budget: $8,000)
  ├── Development MG
  │     ├── Sub: dev (Budget: $2,000)
  │     └── Sub: sandbox (Budget: $500)
  └── Platform MG
        └── Sub: shared-services (Budget: $5,000)

Cost Management supports viewing costs across all subscriptions under a management group — useful for finance and engineering leadership.

Key Takeaways

• Cost Analysis is your main tool — filter by service, resource group, tag, and time range.
• Budgets with alert thresholds catch spending surprises before the bill arrives.
• Tags (enforced via policy) are the foundation of cost attribution — require Environment, Team, and CostCenter.
• Advisor automatically finds right-sizing opportunities, idle resources, and reservation candidates.
• Reservations save 30–65% for steady-state workloads; Savings Plans offer flexibility.
• Spot VMs offer up to 90% discount for fault-tolerant workloads.
• Common traps: forgotten dev VMs, oversized resources, unattached disks, premium storage in dev.