Build sole-tenant node templates, node groups with autoscaling, and instance affinity configurations.
Last verified: May 2026
Build sole-tenant node templates, node groups with autoscaling, and instance affinity configurations.
Required Fields
nodeTemplate.namenodeTemplate.nodeTypenodeTemplate.regionnodeGroup.namenodeGroup.zoneOutput will appear here...Sole-tenant nodes give you dedicated physical Compute Engine servers that host only your project's VMs, providing hardware isolation for compliance, licensing, and performance requirements. They are essential for workloads governed by regulations like HIPAA, PCI-DSS, or financial services mandates that prohibit multi-tenant hardware. This builder helps you configure sole-tenant node templates with specific machine types, affinity labels, CPU overcommit settings, and maintenance policies, generating the configuration for deployment.
Your team has a HIPAA-regulated healthcare app that requires hardware isolation. The builder generates: a sole-tenant node template using n2-node-80-640 (80 vCPU, 640 GB RAM physical server), a node group with 3 nodes for high availability, MIGRATE_WITHIN_NODE_GROUP maintenance policy (migrates VMs between dedicated nodes during host maintenance, never to multi-tenant hardware). Cost: ~$8K/month for the 3 nodes, supports up to ~30 production VMs across them. Compliance audit closed; HIPAA hardware isolation requirement met.
Sole-tenant nodes are a niche feature — only use them when you have HARD compliance or licensing requirements. The cost premium is significant (you pay for whole nodes regardless of VM density), and Google's standard infrastructure already provides strong tenant isolation for non-regulated workloads.
BYOL (Bring Your Own License) on sole-tenant for Windows Server / Oracle Database can be dramatically cheaper than license-included pricing. Validate licensing terms with vendors before assuming this works — some software licenses prohibit running in 'public cloud' even on dedicated hardware.
CPU overcommit (placing more vCPUs than physical cores) makes sense ONLY for workloads with low average CPU. Overcommit ratio 2x is typical for dev/test. For production, run at 1:1 — overcommit causes CPU contention and noisy-neighbor effects that defeat the point of dedicated hardware.
The builder constructs sole-tenant node template + node group resources: node template (machine type for the underlying server, server binding type, optional CPU overcommit ratio, optional accelerator config), node group (template reference, target node count, autoscaling policy, maintenance policy: DEFAULT or RESTART_IN_PLACE or MIGRATE_WITHIN_NODE_GROUP). Output is generated as gcloud compute sole-tenancy commands and Terraform google_compute_node_template + google_compute_node_group resources.
Was this tool helpful?
Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.