What was important yesterday may not be today.
As you learn by running your workloads in production, expect changes. Your configuration, business needs, workflows, and even your team may evolve. You might need to adjust how you develop and deploy your software. External factors can also change, such as the cloud platform, its resources, or your contractual agreements.

Monitor the impact of changes on your costs.
Regularly check whether your return on investment (ROI) is moving in the right direction, and adjust your goals or requirements if necessary.

Example Scenario

Contoso Air provides a baggage tracking solution for airlines. The workload is hosted on Azure, runs on Azure Kubernetes Service (AKS), uses Azure Cosmos DB for the database, and Azure Event Hubs for messaging. It is deployed in the West US and East US regions.

Track and monitor your spending

Use a cost tracking system to regularly review your expenses related to resources, data, and support. If some resources are underutilized, consider stopping, replacing, or reorganizing them for greater efficiency.

Understanding where your money goes is the first step to controlling it.

By tagging resources, categorizing expenses, and setting up alerts, you can track costs by teams, services, and environments.

This helps you:

• Quickly detect unexpected charges
• Support showback or chargeback models
• Make smarter decisions on where to invest or cut

Contoso’s Challenge

The team has always stayed under budget, so cost reduction wasn’t a priority.

But next year, they plan to improve reliability, which will increase Azure costs. This could exceed the budget, so they consider requesting an increase.

Approach and Results

Before asking for more budget, the team closely reviews its current Azure and support costs.

They discover:

• Unused virtual machines linked to an old build system
• Old data in Azure Storage that could be moved to a cheaper tier
• A support contract including unused consulting hours

The team:

• Deletes unused VMs
• Moves data to Archive storage
• Makes better use of included consulting services

They add a recurring task to review and optimize costs regularly.

Continuously adjust your workload

Continuously adapt architectural decisions, resources, code, and workflows based on ROI data.

Cloud environments evolve, and your architecture should too.

Regularly review:

• Metrics
• Performance
• Billing
• Feature usage

Even small adjustments can generate big savings.

Contoso’s Challenge

The team, always under budget, never explored other approaches. They focused on developing new features.

But after discovering waste, they decide to dig deeper.

Approach and Results

The team finds:

• Too many resources allocated to low-priority workflows
• Ability to reduce throughput without affecting performance
• Opportunity to switch to a queuing system to smooth the load

They also discover that a new component in their compute SKU can replace authentication code, reducing maintenance.

Continuously optimize your cloud environment

Make it a habit to regularly check for unused resources or obsolete data and remove them.

These elements can silently accumulate costs.

Deleting unused resources and unnecessary data frees up budget for more important tasks.

Contoso’s Challenge

The team created several temporary environments for testing and performance experiments but never cleaned them up.

They discover:

• Inactive Event Hubs namespaces and Cosmos DB containers
• Old data in hot access tiers, unnecessary for operations or compliance

Approach and Results

The team sets up:

• A monthly cleanup routine
• Tags with expiration dates
• Usage metrics analysis to identify inactive services

They:

• Delete unused AKS node pools
• Remove inactive Event Hubs
• Consolidate Cosmos DB containers

For historical data, they apply lifecycle policies to archive or automatically delete it.

They also review SKUs and downsize oversized services.

Result: reduced expenses, improved operational efficiency, and a cleaner, easier-to-manage cloud environment.