ResourceQuota for the whole vcluster

One ResourceQuota caps the entire virtual cluster. vcluster renders policies.resourceQuota as a real ResourceQuota in the host namespace, so the sum of every pod across every virtual namespace is bounded by one object.

1. enable the quota and set the ceilings.

   policies:
     resourceQuota:
       enabled: true
       quota:
         requests.cpu: "4"
         requests.memory: 8Gi
         count/pods: "20"
   

2. enable a LimitRange too. A CPU or memory quota only admits a pod that declares requests; a pod with no requests is rejected by the quota. The LimitRange supplies default requests so those pods still count against the quota instead of being denied.

   policies:
     limitRange:
       enabled: true
       default:
         cpu: "1"
         memory: 512Mi
         ephemeral-storage: 8Gi
       defaultRequest:
         cpu: 100m
         memory: 128Mi
         ephemeral-storage: 3Gi
   

3. apply with Helm.

   helm upgrade --install my-vcluster vcluster \
     --repo https://charts.loft.sh \
     --version 0.37.2 \
     -n my-vcluster --create-namespace \
     -f values.yaml
   

4. both objects land in the host namespace.

   $ kubectl -n my-vcluster get resourcequota,limitrange
   NAME                            REQUEST                                              LIMIT
   resourcequota/vc-my-vcluster   requests.cpu: 0/4, requests.memory: 0/8Gi, ...
   
   NAME                        CREATED AT
   limitrange/vc-my-vcluster   2026-10-10T...
   

5. watch the ephemeral-storage default on small nodes. vcluster's default LimitRange requests 3Gi ephemeral-storage per container. On a node with little disk, a handful of pods can exhaust it and go Pending. Lower defaultRequest.ephemeral-storage if that bites.

   policies:
     limitRange:
       defaultRequest:
         ephemeral-storage: 256Mi
   

Set limitRange.enabled: auto to turn the LimitRange on automatically whenever the ResourceQuota is on.

Manifests

values.yaml

policies:
  resourceQuota:
    enabled: true
    quota:
      requests.cpu: "4"
      requests.memory: 8Gi
      count/pods: "20"
  limitRange:
    enabled: true
    default:
      cpu: "1"
      memory: 512Mi
      ephemeral-storage: 8Gi
    defaultRequest:
      cpu: 100m
      memory: 128Mi
      ephemeral-storage: 3Gi

https://www.vcluster.com/docs/vcluster/configure/vcluster-yaml/policies/resource-quota https://www.vcluster.com/docs/vcluster/configure/vcluster-yaml/policies/limit-range