Add partial GPU allocation support

[UdayKemaAviz]

This pull request introduces a new Terraform resource for managing partial GPU allocations for tenants, along with supporting changes to the provider and example configuration. The main focus is adding the fabricapi_gpu_allocations resource, which enables users to allocate or deallocate specific GPUs on compute nodes for a tenant. Additional changes improve the provider's usability and safety, and update the example and documentation files.

New GPU Allocations Resource:

• Added a new resource fabricapi_gpu_allocations to the provider, allowing users to manage GPU allocations for a tenant via the Fabric API. This includes resource schema, CRUD logic, and integration with the provider. [1] [2]

API Client and Data Model Enhancements:

• Implemented new data types (GPUOperation, ServerGPUs, GPUAllocationRequest, GPUAllocationResponse) and the ModifyGPUAllocations method in the API client to support the new GPU allocation API endpoint.

Example and Documentation Updates:

• Added an example Terraform configuration (main.tf, variables.tf, terraform.tfvars.example) demonstrating how to use the new GPU allocations resource. [1] [2] [3]

Provider and Resource Improvements:

• Registered the new GPU allocations resource in the provider.
• Updated the tenant_servers resource to require replacement when tenant_name or fabric_name changes, and improved pre-allocation conflict checks to only run for non-shared tenants. [1] [2] [3] [4]

These changes collectively enable fine-grained GPU management for tenants and improve resource lifecycle handling in the provider.

[Copilot]

Pull request overview

This PR adds a new Terraform resource to perform partial GPU allocation/deallocation for a tenant via the Fabric API, plus supporting client types/methods and an example configuration. It also tightens lifecycle semantics in the existing tenant_servers resource and registers the new resource in the provider.

Changes:

• Added fabricapi_gpu_allocations resource with schema and CRUD behavior targeting POST /fabrics/{fabric}/tenants/{tenant}/gpuAllocations.
• Extended the API client with GPU allocation request/response types and a ModifyGPUAllocations method.
• Updated provider registration and added a runnable example for the new resource; improved tenant_servers replacement behavior and conflict checks.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
internal/provider/tenant_servers_resource.go	Requires replacement on tenant/fabric changes; skips server conflict checks for shared tenants.
internal/provider/provider.go	Registers the new fabricapi_gpu_allocations resource.
internal/provider/gpu_allocations_resource.go	Implements the new GPU allocations Terraform resource.
internal/provider/client.go	Adds GPU allocation request/response types and API client method.
examples/gpu-allocations/main.tf	New example usage of fabricapi_gpu_allocations.
examples/gpu-allocations/variables.tf	Inputs for the GPU allocations example.
examples/gpu-allocations/terraform.tfvars.example	Sample values for the GPU allocations example.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.