Metrics Reference

Metric	Description
k8s_namespace_cost_usd	Aggregated cost per namespace (CPU + memory).
k8s_workload_cost_usd	Cost per Deployment/StatefulSet.
k8s_pod_idle_percentage	Percent of spend attributed to unused requests.
ecs_service_cost_usd	EC2/Fargate cost per ECS service.
ecs_task_runtime_seconds	Lifetime per task, useful for batching workloads.
tenant_cost_usd	Spend per cost.clustercost.io/tenant label.
savings_opportunity_usd	Modeled savings for idle or over-sized workloads.

All metrics are exposed via the dashboard REST API (/api/v1/cost/...) and can be forwarded to Prometheus if you need to integrate with your existing observability stack.