A machine type is a predefined virtual-machine shape that allocates a specific ratio of vCPUs and memory to an instance. GCP calls it a machine type; AWS uses instance type, and Azure uses size. section 12.1.1
Machine types turn resource sizing into a controllable infrastructure choice. A server configured with more vCPUs than its workload needs can increase the bill without improving use, so infrastructure-as-code tests can enforce a maximum. section 12.1.1
A practical control workflow is:
- Inspect observed vCPU and memory usage to identify a suitable capacity limit. section 12.1.1
- Query the provider catalog or API for current vCPU and memory values instead of maintaining a complete local catalog. section 12.1.1
Capacity and cost guardrails
For a production environment capped at 32 vCPUs, a policy test rejects an oversized server after retrieving the selected machine type’s current vCPU count from the GCP API. A separate estimate can price five n2d-standard-16 servers and compare the result with a monthly budget. section 12.1.1 section 12.1.2

