Tag coverage

The Tags view (in the sidebar under Insights) shows your tag coverage: how much of your spend carries tags, summarized in the headline Coverage card. Accurate allocation depends on tags. If the tag keys you rely on — team, environment, cost-center — are missing from a chunk of spend, that spend can’t be attributed, and your chargeback numbers drift. Tag coverage measures the gap so you can close it.

Availability

The Tags view appears in the sidebar only when a cloud provider — AWS, Azure, or Google Cloud — is connected. Tag coverage measures tagging discipline on cloud resources, so an organization whose only sources are AI platforms (Anthropic, OpenAI) or Snowflake doesn’t see the view at all — see Tags on non-cloud sources for what those sources offer instead.

The headline Coverage number — the share of your spend on resources carrying any tag — is available on every plan, including Free. The detailed per-tag-key breakdown on this page is a paid feature (Starter, Growth, and Scale); on the Free plan the view is visible but that detail is withheld until you upgrade. See plans & billing.

What it measures

The headline Coverage number is the share of spend on resources that carry any tag at all — even a bare Name=web-1 counts as covered. It splits your spend into tagged vs. untagged so you can see, at a glance, how much of your bill carries some tag and how much is flying blind. Because a single tag qualifies, treat the headline as an upper bound: use the per-key breakdown below to see coverage for the specific keys your allocation actually relies on.

Coverage spans your cloud providers — AWS tags, Azure tags, and Google Cloud labels are all counted, in one combined tag namespace (a key like environment used on two clouds appears as one row). Spend from AI providers and data warehouses is deliberately left out of the denominator: tag governance doesn’t apply to it, so it would only dilute the number.

Azure’s built-in keys don’t inflate the headline

Azure spend arrives with resource_group, invoice_section, and cost_center attached to every row (see Connect Azure). Those keys are useful — they show up in the per-key table and are filterable anywhere tags are — but they don’t count as “tagged” in the headline split. Otherwise Azure would always read ~100% covered regardless of your actual tagging discipline.

When your view spans more than one cloud integration, a Coverage by integration table also rolls the split up per connected source, ranked by spend — so you can spot which cloud or billing account is dragging coverage down before you drill into its tag keys.

Scoping and filters

Two controls narrow the view, and they compose:

• Scope (the picker at the top of every dashboard) chooses which connected sources you’re looking at — the whole organization, one provider, or specific integrations.
• In-page filters refine within that scope:
  • Date Range — defaults to This Month (month to date); switch to the last 3, 6, or 12 calendar months, or a custom number of months, to see coverage over a longer window. History is bounded by your plan’s data retention.
  • Accounts — limit coverage to specific linked accounts, Azure subscriptions, or GCP projects. Pick accounts, then press Apply.

Every number on the page — the headline, the per-integration roll-up, the per-key table, and untagged-by-service — honors the same scope and filters.

Not the same as Effective Coverage %

Don’t confuse tag coverage with Effective Coverage % — a metric that combines commitment coverage and utilization. Both use the word “coverage,” but tag coverage is about tags and attribution, while Effective Coverage % is about commitment efficiency. Both are live in the product.

Coverage by tag key

Coverage is reported per tag key, because a resource can be well tagged for one purpose and untagged for another. You might have environment on 95% of spend but cost-center on only 60% — and it’s the low key that undermines your chargeback. Reviewing coverage key by key tells you exactly where to focus.

Keys that mean the same thing are counted together — env and Environment fold into one row, and technical prefixes like user_ are stripped — so the table shows cleaned-up display names, not the raw keys stored on your cost data. When you turn a coverage gap into an allocation rule, pick the raw key from the rule editor’s selector rather than copying a display name from this table.

Untagged spend

Tag coverage also surfaces the services with the most untagged spend — the biggest single wins. Fixing tagging on one high-spend, poorly tagged service often moves your coverage more than chasing dozens of small resources.

Improving coverage

• Fix it at the source — add the missing tags to the resources, so future spend is attributed automatically.
• Bridge it with rules — where you can’t retag, use allocation rules (virtual tags) to assign the untagged spend to the right cost center.

Re-check the tagged-vs-untagged split after a tagging push to confirm your policy is taking hold. The view defaults to the current billing month to date, so new tags show up as soon as the next cost data lands; switch the Date Range to a longer window to compare coverage before and after the push.

Tags on non-cloud sources

Snowflake, Anthropic, and OpenAI spend doesn’t participate in tag coverage: those platforms have no user-defined resource tags to govern, so a coverage percentage would be meaningless there. Their spend still carries tag-like dimensions that work everywhere else tags do — the report builder’s tag dimensions, allocation rules, and budget filters:

• Snowflake — warehouse activity is tagged warehouse, so you can pivot credits and estimated spend per warehouse in Reports (the Warehouse dimension) and allocate per warehouse with warehouse tag rules. Spend rows also carry billing dimensions like rating_type, billing_type, and balance_source. Tags you define inside Snowflake (CREATE TAG / ALTER … SET TAG) are not currently imported — see Connect Snowflake.
• Anthropic and OpenAI — rows carry dimensions such as service tier (including batch), context window, and cost type, and spend is already organized by workspace or project as sub-accounts — so attribution works out of the box without any tagging effort.

Related

• Turn coverage gaps into allocation with allocation rules.
• See the payoff on the Allocation page.