As the SDDC expands, you can add compute clusters.

To support Kubernetes workloads deployed by VMware Enterprise PKS, a series of discrete compute clusters will be deployed. You add compute clusters to increase the available resources for Kubernetes workloads or to increase application availability across additional VMware Enterprise PKS Availability Zones. The design decisions determine the number of vSphere hosts, vSphere HA settings, and other characteristics of each compute cluster.

Decision ID

Design Decision

Design Justification

Design Implication


Create three compute vSphere clusters to which Kubernetes clusters are deployed.

Segmenting user workloads from management provides for security, performance, and availability of the solution.

Multiple discrete compute vSphere Clusters for deploying Kubernetes clusters adds to the performance and availability of the solution. Segmenting VMware Enterprise PKS Availability Zones based on vSphere Resource Pools does neither of those.

Workload deployment has to take the physically disparate Availability Zones into account.


Configure Admission Control for one ESXi host failure and percentage-based failover capacity.

vSphere HA protects tenant workloads in the event of an ESXi host failure. vSphere HA powers on virtual machines from the failed ESXi hosts on any remaining ESXi hosts.

Only a single ESXi host failure is tolerated before potential resource contention.


Deploy NFS-based storage for compute cluster use.

NFS storage presented to each ESXi host in all compute clusters ensures all Kubernetes volumes are available in any Availability Zone.

You will have to manage and monitor the external NFS storage platform and any network connectivity.


Create each compute vSphere cluster with a minimum of three ESXi hosts.

Allocating three ESXi hosts provides full redundancy within the cluster.

With the use of NFS for storage of VMs and Kubernetes volumes, you can start with as few as three nodes per compute cluster.


Use the shared edge and compute cluster-based Host Profile.

Using host profiles simplifies configuration of ESXi hosts and ensures settings are uniform across the cluster.

After you add the ESXi hosts as Transport Nodes, the host profile are no longer usable and must be removed from the cluster.