Cost Optimisation Pillar

The Cost Optimisation pillar of the Azure Well-Architected Framework focuses on maximising the value of your cloud spend. It is not about spending as little as possible — it is about ensuring that every pound or dollar you spend delivers business value. Over-provisioning wastes money; under-provisioning degrades performance. The goal is to find the right balance.

---

Design Principles

Plan and Estimate

Before deploying any workload, estimate its cost using the Azure Pricing Calculator. Understand the pricing model for every service you plan to use. Many teams are surprised by their first Azure bill because they did not account for data transfer, storage transactions, or log ingestion costs.

Monitor and Optimise Continuously

Cost optimisation is not a one-time exercise. Cloud costs change as your workload evolves, as Azure introduces new pricing models, and as usage patterns shift. Implement continuous monitoring and review processes.

Maximise Utilisation

Cloud resources that sit idle are wasted money. Right-size your resources, shut down what you do not need, and use autoscaling to match capacity to demand.

Choose the Right Pricing Model

Azure offers multiple pricing models for the same services. Choosing the right one can save 40-90% compared to pay-as-you-go pricing.

---

Understanding Azure Costs

Cost Components

Azure costs typically fall into several categories:

Component	Description	Example
Compute	VMs, App Service, Functions, Containers	VM running 24/7
Storage	Blob, Disk, File, Database storage	1 TB of Blob Storage
Networking	Data transfer, Load Balancer, VPN Gateway	Egress data leaving Azure
PaaS Services	Managed databases, AI, analytics	Azure SQL Database DTUs
Licences	Windows Server, SQL Server	Included or bring your own

Key Pricing Concepts

• Pay-as-you-go — billed per second, per hour, or per transaction with no upfront commitment
• Reserved Instances — commit to 1 or 3 years for up to 72% discount on compute
• Savings Plans — commit to a fixed hourly spend for up to 65% discount across compute services
• Spot VMs — use spare Azure capacity at up to 90% discount (can be evicted at any time)
• Azure Hybrid Benefit — reuse existing Windows Server or SQL Server licences for up to 40% savings

---

Cost Optimisation Strategies

Right-Sizing

Right-sizing means using the smallest resource tier that meets your performance requirements. Azure Advisor analyses your resource utilisation and recommends smaller sizes when resources are underutilised.

Example: If a Standard_D4s_v5 VM (4 vCPUs, 16 GB RAM) consistently uses less than 20% CPU and 4 GB RAM, downsize to a Standard_D2s_v5 (2 vCPUs, 8 GB RAM) and cut your compute cost in half.

Shutting Down Idle Resources

Development and testing environments often run 24/7 but are only used during business hours. Use auto-shutdown schedules or Azure DevTest Labs to stop resources outside working hours:

• A VM running 10 hours/day instead of 24 reduces compute cost by approximately 58%
• Use Azure Automation or Azure Functions to start/stop resources on a schedule

Reserved Instances and Savings Plans

For production workloads with predictable, steady usage:

Commitment	Typical Savings	Best For
1-year Reserved Instance	Up to 40%	Stable production VMs and databases
3-year Reserved Instance	Up to 72%	Long-term, predictable workloads
Savings Plan	Up to 65%	Mixed compute usage across services