Cloudeka Resource

The live billing mirror of a running Pod. For every Pod you run, the platform keeps one of these showing its compute footprint and current hourly charge. It's created automatically when your Pod is admitted and removed automatically when the Pod is gone — you don't manage it directly.

kubectl get container -n <your-namespace>
kubectl get container -n <your-namespace> -o yaml

Field	Meaning
status.kind	Workload kind (e.g. pod/deployment)
status.cpu	CPU currently in use
status.ram	RAM currently in use
status.gpu.<model>	GPU count in use, per model
status.nvaie	Whether NVAIE pricing applies
status.chargePerHour	Current hourly cost of this workload
status.status	Mirrors the Pod's phase: Running, Pending, or other

Find your most expensive running Pods:

kubectl get container -n <ns> --sort-by=.status.chargePerHour

Client-side sort; not all kubectl versions support this flag for CRDs — otherwise use -o yaml and inspect manually.

Confirm a Pod's cost matches expectations before leaving it running long-term.

This object is fully system-managed — it follows the Pod it mirrors. Deleting it directly does not stop billing or delete the Pod; delete the Pod instead.