Cost Query Engine
Build custom cost queries with multi-dimensional filters to find exactly where spend is going.
Cost Query Engine is fully functional but the interface may evolve based on feedback.
Overview
Cost Query Engine
Cost Query Engine lets you slice and filter cluster costs across multiple dimensions. Build queries by combining filters, view results by resource type, and save filter combinations as dashboards for quick access.
Access: Cost Query Engine (top-level navigation)
Filters
Add one or more filters to narrow results. Click "Add Filter", pick a filter type, configure values, then run the query.
Available filter types:
- Cluster - Scope to specific clusters
- Namespace - Limit to specific namespaces
- Team - Filter by team label
- Owner - Filter by owner label
- Department - Filter by department label
- Workload Type - Deployment, StatefulSet, DaemonSet, Job
- Instance Type - EC2/GCE instance type (e.g. m5.xlarge)
- Node Role - Filter by node role
- Storage Class - Filter by PV storage class (gp2, gp3, io1)
- Labels - Filter by arbitrary Kubernetes label key-value pairs
Active filters display as chips in the filter bar. Remove a filter by clicking the X on its chip.
Resource Tabs
Results are grouped by resource type. Switch tabs to view different resource categories:
- All - Every resource type in a single view
- Workloads - Deployments, StatefulSets, DaemonSets, Jobs
- Namespaces - Namespace-level cost aggregation
- Nodes - Individual node costs and utilization
- Volumes - Persistent volume costs and capacity
Each tab shows a paginated table with columns for name, cluster, namespace, type, cost, and efficiency. Click a resource name to open its detail view. Sort by cost or name.
KPI Summary
Above the tabs, four summary cards display aggregate metrics for the current query:
- Hourly Cost - Combined hourly cost of filtered resources
- Daily Cost - Projected daily cost
- Monthly Cost - Projected monthly cost
- Total Resources - Count of resources matching the query
A cost trend chart shows daily costs over the selected timeframe (7d, 14d, or 30d).
Saved Dashboards
Save a filter combination as a named dashboard for quick reuse. Saved dashboards are stored on the server and shared with your organization.
Saving dashboards requires organization admin permissions. All org members can load and view saved dashboards.
Save a dashboard:
- Configure desired filters
- Click "Save Dashboard"
- Enter a name
- Dashboard is saved to your organization
Load a saved dashboard:
- Click the dashboard dropdown
- Select from the list
- Filters are applied automatically
- Click "Run Query" to execute
Delete a dashboard:
- Select a saved dashboard, then use the delete option