48.4k Log in Get started
Dashboard

What goes in a Kubernetes cost dashboard?

A Kubernetes cost dashboard breaks cluster spend down to namespaces, workloads, and teams — including the idle capacity and unallocated shared cost that most views quietly omit. Build it on daily allocation rollups from Kubecost or OpenCost landed in your warehouse.

For: Platform engineering, team leads, and FinOps. Refresh: daily.Source: modeled billing and cost tables in a Metabase-connected database.

What does this dashboard look like?

The layout runs from cluster cost and idle capacity, through namespace, workload and team breakdowns, to request-versus-usage efficiency and the reconciliation to the cloud bill.

Kubernetes cost dashboard in Metabase showing cluster cost, namespace and team splits, idle capacity, and efficiency.
An example kubernetes cost dashboard in Metabase, built from modeled billing and cost tables. Figures are illustrative.

Which cards belong on this dashboard?

  • Cluster cost this month and month-over-month change
  • Cost by namespace, top 15
  • Cost by workload within a selected namespace
  • Idle cost share and its trend
  • CPU vs. memory vs. storage vs. network cost split
  • Efficiency by namespace (used vs. requested)
  • Cost by team label, with the unlabeled bucket visible
  • Cluster cost reconciled against the cloud bill

What data does this dashboard need?

  • Daily allocation rollups per namespace and workload (Kubecost/OpenCost Allocation API)
  • Cost components: CPU, memory, storage, network, GPU where present
  • Idle and unallocated cost broken out explicitly
  • Efficiency (usage vs. request) per workload
  • Label-to-team mapping for allocation views

How do you build it?

  1. Land billing exports or cost-platform data in a database and keep raw line items, tags, timestamps, and pricing models.
  2. Build rollup models at the grain this dashboard needs, with cost conventions (amortized, unblended, net) as explicit columns.
  3. Create one saved question per card and certify the shared models and definitions.
  4. Add dashboard filters for Date range, cluster, namespace, team label, environment.
  5. Show data freshness — billing exports lag, and every viewer should see by how much.

Example card SQL

Namespace cost with idle share, trailing 30 daysPostgreSQL
SELECT
  namespace,
  ROUND(SUM(cpu_cost + ram_cost + pv_cost + network_cost), 2) AS direct_cost,
  ROUND(SUM(idle_cost), 2) AS idle_cost,
  ROUND(
    100.0 * SUM(idle_cost)
    / NULLIF(SUM(cpu_cost + ram_cost + pv_cost + network_cost + idle_cost), 0),
    1
  ) AS idle_share_pct
FROM k8s_allocation_rollups
WHERE window_start >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY namespace
ORDER BY direct_cost DESC;

Dashboards

Cloud spend overview dashboardCost allocation by team and tag dashboardCost per customer dashboardCommitment coverage dashboardCost anomaly dashboard

Integrations

Cloud cost & FinOps integrationsAWS Billing + MetabaseGoogle Cloud Billing + MetabaseVantage + MetabaseKubecost + MetabaseCloudZero + Metabase

Metrics

Total cloud spendCost per customerCost growth rateCommitment coverage rateEffective savings rateUntagged spend rateBudget burn rateCost per service

Analytics

Cloud cost analytics

FAQ

Why doesn't my namespace total match the cloud bill?
Allocation distributes measured usage over assumptions — node pricing, shared resources, and idle capacity. Add idle and unallocated cost back in and reconcile monthly against the provider's billing export; a small gap is normal, a growing one isn't.
Do I need Kubecost, or is OpenCost enough?
Both expose the same Allocation API shape this dashboard is built on. OpenCost (CNCF) covers single-cluster allocation and is free; Kubecost adds multi-cluster aggregation, savings insights, and enterprise support. Start with either — the warehouse model doesn't change.