Skip to content

Cost Explorer

Cost Explorer is the analytical companion to the Financial portal. Where Financial is about invoices and payments, Cost Explorer is about understanding spend — by service, by tag, by project, over time.

Overview

Cost Explorer Cost Explorer — chart on top, table below. Slice and dice with filters and group-by.

Open from Financial → Explore or directly at /project/<id>/account/financial/explorer.

Top of the page: - Time range picker — Last 7d / Last 30d / This month / Last month / Custom - Granularity — Hourly / Daily / Monthly - Group by — Service / Region / Tag / Project / Resource

Main panel: - Chart — stacked bars (default) or area chart - Table — every line item that contributes to the chart, sortable by spend, by name, by trend %

Right rail: - Filters — region, service, tag, resource ID, account - Saved views — your bookmarked queries - Export — CSV / JSON / scheduled email digest

Administration

Data freshness

Cost Explorer data updates hourly for current-month cost and daily for prior-month and forecast. Real-time per-API-call cost is not exposed (use the underlying meters via API if you need second-level granularity).

Retention

View Retention
Hourly granularity 30 days
Daily granularity 18 months
Monthly granularity 84 months (7 years, NBR-aligned)

Permissions

Role Cost Explorer access
Org Owner / Admin / Billing Full, all projects
Project Admin Full, this project only
Project Member Read-only, this project only
Project Read-only Read-only, this project only

Tag visibility

For tag-based drill-down to be useful, tagging must be consistent. The Cost Explorer surfaces:

  • Coverage warning — % of resources missing required tags (per project tagging policy)
  • Untagged spend — what's spent on resources with no cost-center (or whatever you've made required)

Higher untagged-spend means lower chargeback accuracy. Aim for >95% coverage.

Operation

Common queries

Question Setup
"What did we spend on compute last month?" Time = Last month · Group by Service · Filter Service = Servers + Volumes
"Top 10 most expensive resources this month" Time = This month · Group by Resource · Sort by spend desc · head 10
"Compare cost-center cc-101 month-over-month" Time = Last 6 months · Group by Tag (cost-center) · Filter cost-center = cc-101
"Which region am I spending most in?" Time = Last 30d · Group by Region
"What's the trend on Object Storage?" Time = Last 6 months · Granularity = Monthly · Filter Service = Object Storage
"Untagged spend this month" Time = This month · Filter Tag (cost-center) = (none)
"Hourly burn during the load test" Time = Custom (the hour) · Granularity = Hourly · Filter Project = mybank-loadtest

Saving views

A useful query → Save → name it → it's bookmarked in the right rail. Saved views are per-user; share by URL (the URL captures all the filter state).

Scheduled reports

Save → Schedule to email a CSV of the current view daily / weekly / monthly to a distribution list. Useful for finance teams who want passive visibility without logging into Cloud Digit.

Exporting

Format Use
CSV Excel, Sheets, finance-team analysis
JSON Scripting, BI ingestion
Scheduled email Recurring report
Webhook Push to your data lake / BI on a schedule

Comparing periods

Hold Shift in the time-range picker to add a comparison range. The chart overlays the comparison; the table adds a "% change" column. Useful for month-over-month or year-over-year analysis.

Anomaly view

Cost Explorer → Anomalies surfaces line items where the current burn rate is materially different from the same range one period earlier. Tunable threshold (default 2×). Each anomaly links back to the resource that caused it.

Common patterns

Chargeback / showback

Goal: produce a per-cost-center bill for internal accounting.

  1. Tagging policy in every project requires cost-center tag at create-time
  2. Cost Explorer view: Group by Tag (cost-center), This month
  3. Export as CSV
  4. Hand to finance — they upload to the chargeback system

Coverage warnings tell you when tagging discipline is slipping.

Cost optimization

Goal: find waste.

  1. Top 20 most expensive resources view
  2. Cross-check with data-classification tag — Public data on Provisioned IOPS is over-engineered
  3. Cross-check with last-modified — large volumes / images that haven't been touched in 90 days are candidates for deletion
  4. Use Anomalies to find regressions

Forecast vs actual

Goal: trust the budget.

  1. Cost Explorer → Compare → This month vs Forecast — see early-month variance
  2. If actual is materially over forecast, drill in to find the cause
  3. Adjust either spend or budget

Troubleshooting

Symptom Likely cause Fix
"No data" for a recent time range Hourly aggregator hasn't run yet Wait up to 60 min; or use Daily granularity which is closer to real-time
Total in Cost Explorer doesn't match invoice Cost Explorer is approximate (hourly snapshots); invoice is authoritative monthly Use invoice for reconciliation, Cost Explorer for analysis
Tag filter shows fewer resources than expected Some resources missing the tag Check tagging-coverage warning; back-fill tags
Forecast wildly off Recent burst skewed the linear projection Switch to median-based forecast in view settings
CSV export missing rows Time range > 90 days at hourly granularity hits row limit Switch to Daily granularity or split the export
Saved view broken after rename of resource Filter pinned to old resource ID Edit the saved view's filter to use the new ID, or filter by name pattern instead