Kubernetes cost, allocated
Allocate Kubernetes spend down to cluster, namespace, and pod, and surface over-provisioned workloads that are quietly burning money.
A shared cluster hides who owns what. Node bills don’t map to teams or services, and over-requested pods waste capacity that never shows up as a line item.
Built to do the work, not just show it
Cluster → namespace → pod
Break cluster cost down to the workloads actually consuming it.
Right-sizing signals
Flag over-provisioned pods whose requests dwarf real usage.
Team attribution
Map namespaces to owners so cost has an address.
Multi-cluster
Aggregate across clusters in the workspace.
From signal to result
Ingest
Cluster + workload metrics are collected.
Allocate
Cost is distributed to namespaces and pods.
Right-size
Over-provisioned workloads surface as findings.
- Over-provisioned-pod findings flow into the same remediation funnel.
- Allocation is per cluster and rolls into the portfolio.
- Works alongside AWS + GCP cost, not as a separate silo.
Questions
How is pod cost computed?
From cluster cost distributed by workload resource usage down to the pod.
Can it recommend right-sizing?
Yes — over-provisioned pods surface as findings you can act on.