Azure FinOps: a practical guide to controlling Azure costs across your desktop estate

Phase one: inform (visibility and allocation)

You cannot improve what you cannot see, and you cannot hold anyone accountable for costs they cannot attribute to their own work. The inform phase builds that foundation.

Microsoft Cost Management and Cost Analysis

Unexpected spend shows up first in Cost Analysis. It slices costs by subscription or resource group, then by tags and services. Anomaly detection sits inside Cost Analysis's smart views at the subscription level, comparing what you forecasted against the day's actual usage over the last 60 days. The anomaly signal gives you an early warning when someone spins up something expensive and forgets about it.

For larger estates, the manual view is not enough. Microsoft recommends scheduled exports as the most scalable way to move cost data out of the portal, and those exports can use the FOCUS format: the FinOps Open Cost and Usage Specification, a provider-agnostic schema that makes your cost data portable across clouds.

Cost Management only retains cost data for 13 months, so export it if you need a longer history.

Tagging and cost allocation

Tags supply the fields Cost Management uses for allocation. Azure supports up to 50 tag name-value pairs at each supported scope. Manual tagging drifts when teams are busy.

Two mechanics fix the drift: 

• Tag inheritance automatically copies tags down from a billing profile, resource group, or subscription to the resources underneath it, so you don't have to tag every resource by hand.
• Cost allocation lets you split the cost of a shared resource, like a network gateway or log workspace, across the business units that actually use it. That's the difference between showback, showing teams what they cost, and real chargeback, billing them for it.

To keep tagging consistent, use Azure Policy. Its Modify effect can add, update, or remove tags on a resource automatically at creation, and remediation tasks can fix tags on resources that already exist and don't comply. This is where tagging stops being a spreadsheet exercise and becomes enforced governance.

Structure your Azure hierarchy to match. Per Microsoft's allocation guidance, management groups handle policy assignment and organization-wide reporting, subscriptions give engineering teams room to work, and resource groups set granular deployment boundaries.

Budgets and anomaly alerts

Budgets turn visibility into early warning. You can set them at budget scopes, including subscription, resource group, or management group scope, with up to five thresholds and five email addresses each. Actual-cost alerts fire when spend crosses a threshold. Forecasted alerts warn you before spend is likely to exceed budget, which gives you time to act instead of explain.

Budgets at subscription or resource group scope can call an Action Group to trigger an automated response when spend hits a threshold. Route that into your ITSM tool and a budget breach becomes a ticket rather than a surprise on the invoice.