Azure Cost Management + Billing: Budgets, Analysis, and Resource Tags

Azure Cost Management + Billing is Azure’s dedicated suite for tracking, allocating, and controlling spending — it’s where you view what you’ve spent, set thresholds that trigger alerts, forecast where you’re headed, and export data for outside reporting. Resource tags supply the labels that make all of that reporting meaningful by department, project, or cost center. Together they answer three questions every organization asks about its cloud spend: what did we spend, who spent it, and are we about to overspend.

What Cost Management + Billing actually does

Cost Management + Billing is distinct from tools that sound adjacent but do something else entirely. Azure Monitor tracks performance and health telemetry, not spend. Azure Policy enforces configuration rules. Microsoft Purview handles data governance and cataloging. Cost Management + Billing’s job is narrower and specific: monitor, allocate, and optimize spending across your Azure environment. Its two headline capabilities are cost analysis, for exploring what already happened, and budgets, for getting ahead of what’s about to happen.

Cost analysis: exploring what you’ve already spent

Cost analysis is the self-service reporting view inside Cost Management. It lets you break down and filter historical spending by service, resource, resource group, or tag, over whatever date range you pick — no support ticket required. A finance team that wants a visual breakdown of last quarter’s spend by resource group and service pulls it directly from cost analysis rather than waiting on anyone.

Cost analysis is also scope-flexible. You can view it at the level of an individual resource, a resource group, a subscription, or a management group, and the higher scopes automatically roll up everything beneath them. A holding company with five subsidiary subscriptions organized under one management group gets a single consolidated view of all five simply by viewing cost analysis at the management group scope — no manual combining of five separate reports required.

Cost analysis includes forecasting, too: it projects likely future spend by extrapolating from the usage trend observed in the current billing period, which is what lets budgets warn you before an invoice arrives rather than after.

Budgets: getting warned before you overspend

A budget sets a spending threshold — say, a monthly dollar amount — and configures alert conditions, often set as a percentage of that threshold, that fire when actual or forecasted spend approaches or crosses it. A healthcare organization that wants an automatic email when spending is projected to hit 80% of its monthly allotment configures exactly this: a budget with an alert condition at the 80% forecasted threshold.

It’s worth being precise about what a budget does not do. Crossing a budget threshold sends a notification; it does not raise your spending limit, and critically, it does not stop resources from running or automatically shut anything down. If a team’s spend blows past a 100% alert and their VMs keep running and accumulating charges, that’s expected behavior — budgets notify, they don’t enforce. Stopping resources automatically in response requires pairing the alert with separate automation, such as an action group that triggers a runbook or Logic App.

Cost Management also runs anomaly detection independently of any budget you’ve configured. It compares current spend against historical patterns and can alert you to an unexpected spike even when no threshold has been crossed — useful for catching a runaway resource or a misconfiguration before it becomes a budget problem at all. Where anomaly detection flags unexpected spend after the fact, Azure Advisor’s cost recommendations work proactively, analyzing usage telemetry to suggest resizing or shutting down VMs that are already underutilized.

Resource tags: the metadata that makes reporting possible

A resource tag is a name-value pair of metadata — something like CostCenter: Marketing — applied to a resource, resource group, or subscription. Tags don’t affect how a resource functions or what it costs; they exist purely to organize, categorize, and, most importantly for this topic, report on spending. When a consulting firm has client-project resources scattered across several resource groups within one subscription, a consistent tag such as Project with a distinct value per client lets cost analysis group and total spend by that tag regardless of which resource group each resource physically sits in. That cross-cutting grouping is exactly what tags are built for, and it’s the mechanism behind internal chargeback — attributing incurred cost back to the department or project that generated it.

Two technical details about tags come up often on the exam and in practice. First, tags are not automatically inherited: a resource group tagged Environment: Production does not pass that tag down to a VM created inside it. Each resource needs its own tag unless you enforce inheritance through Azure Policy — which is also how you’d require a CostCenter tag on every new resource so chargeback reports never end up with untagged gaps. Second, Azure caps tagging at 50 name/value pairs per resource, resource group, or subscription; tag names are case-insensitive but tag values preserve case.

Capability	Answers	Key behavior
Cost analysis	What did we spend, broken down by what?	Self-service, scoped from resource up to management group, includes forecasting
Budgets	Are we about to overspend?	Alerts on actual or forecasted thresholds; does not stop resources
Anomaly detection	Did something spend unexpectedly?	Flags deviations from historical patterns automatically, no threshold needed
Scheduled exports	Can finance get this data outside the portal?	Recurring delivery of cost/usage data to a storage account for Excel or Power BI
Resource tags	Whose spend is this?	Name/value metadata enabling grouping and chargeback across resource groups

Scheduled exports for external reporting

When a finance team wants Azure cost data flowing automatically into Power BI or Excel every week without anyone manually downloading anything, the answer is a scheduled export: a recurring job that delivers detailed cost and usage data to an Azure Storage account on a schedule you define. Combine that with consistent department tags and finance gets a fully automated, tag-grouped spending breakdown landing in their reporting pipeline with zero manual steps.

How the AZ-900 exam tests this

• A scenario names a specific reporting need — “notify us before we overspend,” “break down spend by resource group and service,” “get the data into Power BI automatically” — and asks which capability fits; match the need to budgets, cost analysis, or scheduled exports respectively rather than defaulting to the first Cost Management answer you see.
• A question describes a tag applied to a resource group and asks whether resources inside it automatically get that tag — the answer is no, tags don’t cascade without Azure Policy enforcing inheritance.
• A scenario stacks multiple requirements (chargeback across business units, forecasting, and per-subscription overspend alerts) in one question — the exam wants you to recognize that no single capability covers all three; it takes tags plus scoped cost analysis plus per-subscription budgets together.
• A question tests whether you know budgets only alert rather than enforce — watch for a distractor implying a budget automatically deletes or stops resources at 100%.

Quick reference

• Cost analysis explores historical spend, filterable by service, resource group, or tag, at any scope from a single resource up to a management group.
• Budgets alert on spending thresholds (actual or forecasted) but never automatically stop resources — that requires separate automation.
• Anomaly detection flags unusual spend patterns independently of any configured budget.
• Scheduled exports push cost/usage data to a storage account on a recurring basis for tools like Power BI or Excel.
• Resource tags are name/value metadata (50 max per resource, resource group, or subscription) used for cost allocation and chargeback.
• Tags are not automatically inherited from a resource group to the resources inside it — use Azure Policy to enforce that.
• Management group scope in cost analysis rolls up spend across every subscription beneath it into one consolidated view.

Once you’re comfortable with what each capability answers, pair this with what actually drives Azure cost in the first place, how Azure Policy enforces rules like required tags, and the purchasing options that cut compute costs before they ever hit a budget; see the full AZ-900 study guide for how this fits into all three exam domains, then test yourself with AZ-900 practice questions that mix budgets, tags, and cost analysis in the same scenario.