Version: 1.14.0

GPR Templates

This topic describes the REST APIs to manage GPR Templates.

Create a GPR Template

Use this API to create a GPR Template.

Method and URL

POST /api/v1/gpr-template

Parameters

Parameter	Parameter Type	Description	Required
name	String	The name for the GPR template.	Mandatory
clusterName	String	The name of the cluster this template is being created for.	Mandatory
numberOfGPUs	Integer	The number of GPUs that needs to be provisioned.	Mandatory
instanceType	String (as available on the inventory list)	GPU node instance type that you can get from the Inventory list details.	Mandatory
numberOfGPUNodes	Integer	Number of GPU Nodes that you want to request.	Mandatory
priority	String	Priority of the request. You can change the priority of a GPR in the queue. You can select a GPR and increase the priority number (low: 1-100, medium: 101-200, high: 201-300) to move a GPR higher in the queue.	Mandatory
memoryPerGPU	Integer	The memory per GPU is specified using this parameter. The default value is set if you do not pass this parameter.	Mandatory
gpuShape	String (as available on the inventory list)	Name of the GPU type that you can get from the List Inventory details.	Mandatory
exitDuration	Timestamp	The duration for which you want the GPU allocation in the timestamp format, DDHHMM.	Mandatory
idleTimeOutDuration	Timestamp	The duration for which a GPU node can be considered idle before it can be used by another GPR.	Optional
enforceIdleTimeOut	Boolean	If you set the `idleTimeOutDuration`, then the value of this parameter is enabled by default. Set the value to `false` if you do not want it enforce the idle time out.	Optional
enableEviction	Boolean	Enable this option to enable auto-eviction of the low priority GPR.	Optional
requeueOnFailure	Boolean	Enable this option to requeue GPR in case it fails.	Optional

Example Request

curl -X POST --location --globoff --request
'{
    "name":"final-tes1",
    "clusterName":"worker-1",
    "numberOfGPUs":1,
    "instanceType":"n1-highcpu-2",
    "exitDuration":"0d0h1m",
    "numberOfGPUNodes": 1,
    "priority":201,
    "memoryPerGPU":15,
    "gpuShape":"Tesla-T4",
    "enableEviction": true,
    "requeueOnFailure": true,
    "idleTimeOutDuration": "0d0h1m",
     "enforceIdleTimeOut":true
}'

Example Responses

Successful Response

{
  "statusCode": 200,
  "status": "OK",
  "message": "Success",
  "data": {
    "gprTemplateName": "api-test1"
  }
}

Bad Request

{
  "statusCode": 409,
  "status": "UPSTREAM_REQUEST_ERROR",
  "message": "Error while creating a GPR Template",
  "data": null,
  "error": {
    "errorKey": "UPSTREAM_REQUEST_ERROR",
    "message": "Error while creating a GPR Template",
    "data": "Error: gprtemplates.gpr.kubeslice.io \"api-test1\" already exists; Reason: AlreadyExists"
  }
}

Get a GPR Template

Use this API to get a specific GPR template by calling its name.

Method and URL

GET /api/v1/gpr-template?gprTemplateName=<template-name>

Parameters

Parameter	Parameter Type	Description	Required
gprTemplateName	String	The name of the GPR template that you want to retrieve.	Mandatory

Example Request

curl -X GET --location --globoff --request

{{host}}/api/v1/gpr-template?gprTemplateName=vc-test1-template

Example Response

{
  "statusCode": 200,
  "status": "OK",
  "message": "Success",
  "data": {
    "name": "vc-test1-template",
    "clusterName": "worker-1",
    "numberOfGPUs": 1,
    "instanceType": "n1-highcpu-2",
    "exitDuration": "0h12m0s",
    "numberOfGPUNodes": 1,
    "gpuShape": "Tesla-T4",
    "memoryPerGpu": 15,
    "priority": 201,
    "enableEviction": true,
    "requeueOnFailure": true,
    "idleTimeOutDuration": "",
    "enforceIdleTimeOut": false
  }
}

List GPR Templates

Use this API to get the list of all GPR templates.

Method and URL

GET /api/v1/gpr-template/list

Parameters

None

Example Request

curl -X GET --location --globoff --request

{{host}}/api/v1/gpr-template

Example Response

{
  "statusCode": 200,
  "status": "OK",
  "message": "Success",
  "data": {
    "items": [
      {
        "name": "9999",
        "clusterName": "worker-2",
        "numberOfGPUs": 1,
        "instanceType": "a2-highgpu-2g",
        "exitDuration": "75h3m0s",
        "numberOfGPUNodes": 1,
        "gpuShape": "NVIDIA-A100-SXM4-40GB",
        "memoryPerGpu": 40,
        "priority": 201,
        "enableEviction": false,
        "requeueOnFailure": false,
        "idleTimeOutDuration": "",
        "enforceIdleTimeOut": false
      },
      {
        "name": "abc",
        "clusterName": "worker-1",
        "numberOfGPUs": 1,
        "instanceType": "a2-highgpu-2g",
        "exitDuration": "96h0m0s",
        "numberOfGPUNodes": 1,
        "gpuShape": "NVIDIA-A100-SXM4-40GB",
        "memoryPerGpu": 40,
        "priority": 101,
        "enableEviction": true,
        "requeueOnFailure": true,
        "idleTimeOutDuration": "72h0m0s",
        "enforceIdleTimeOut": true
      },
      {
        "name": "auto-gpr",
        "clusterName": "worker-1",
        "numberOfGPUs": 2,
        "instanceType": "n1-highcpu-2",
        "exitDuration": "1h0m0s",
        "numberOfGPUNodes": 1,
        "gpuShape": "n1-highcpu-2",
        "memoryPerGpu": 15,
        "priority": 101,
        "enableEviction": true,
        "requeueOnFailure": true,
        "idleTimeOutDuration": "",
        "enforceIdleTimeOut": false
      }
    ]
  }
}

Update a GPR Template

Use this API to update a specific GPR template.

Method and URL

PUT /api/v1/gpr-template

Parameters

Parameter	Parameter Type	Description	Required
name	String	The name for the GPR template.	Mandatory
clusterName	String	The name of the cluster this template is being updated for.	Mandatory
numberOfGPUs	Integer	The number of GPUs that needs to be provisioned.	Mandatory
instanceType	String (as available on the inventory list)	GPU node instance type that you can get from the Inventory list details.	Mandatory
numberOfGPUNodes	Integer	Number of GPU Nodes that you want to request.	Mandatory
priority	String	Priority of the request. You can change the priority of a GPR in the queue. You can select a GPR and increase the priority number (low: 1-100, medium: 101-200, high: 201-300) to move a GPR higher in the queue.	Mandatory
memoryPerGPU	Integer	The memory per GPU is specified using this parameter. The default value is set if you do not pass this parameter.	Mandatory
gpuShape	String (as available on the inventory list)	Name of the GPU type that you can get from the List Inventory details.	Mandatory
exitDuration	Timestamp	The duration for which you want the GPU allocation in the timestamp format, DDHHMM.	Mandatory
idleTimeOutDuration	Timestamp	The duration for which a GPU node can be considered idle before it can be used by another GPR.	Optional
enforceIdleTimeOut	Boolean	If you set the `idleTimeOutDuration`, then the value of this parameter is enabled by default. Set the value to `false` if you do not want it enforce the idle time out.	Optional
enableEviction	Boolean	Enable this option to enable auto-eviction of the low priority GPR.	Optional
requeueOnFailure	Boolean	Enable this option to requeue GPR in case it fails.	Optional

Example Request

curl -X PUT --location --globoff --request
{{host}}/api/v1/gpr-template

'{
    "name":"312",
    "clusterName":"worker-1",
    "numberOfGPUs":1,
    "instanceType":"n1-highcpu-2",
    "exitDuration":"0d0h1m",
    "numberOfGPUNodes": 1,
    "priority":201,
    "memoryPerGPU":15,
    "gpuShape":"Tesla-T4",
    "enableEviction": true,
    "requeueOnFailure": true,
    "idleTimeOutDuration": "0d0h1m",
    "enforceIdleTimeOut":true
}'

Example Response

{
  "statusCode": 200,
  "status": "OK",
  "message": "Success",
  "data": {}
}

Delete a GPR Template

Use this API to delete a specific GPR template by calling its name.

Method and URL

DELETE /api/v1/gpr-template

Parameters

Parameter	Parameter Type	Description	Required
gprTemplateName	String	The name of the GPR template that you want to delete.	Mandatory

Example Request

curl -X DELETE --location --globoff --request
{{host}}/api/v1/gpr-template

'{
    "gprTemplateName": "api-test1"
}'

Example Responses

Successful Response

{
  "statusCode": 200,
  "status": "OK",
  "message": "Success",
  "data": {}
}

Bad Request

{
  "statusCode": 404,
  "status": "UPSTREAM_REQUEST_ERROR",
  "message": "Error while deleting GPR template",
  "data": null,
  "error": {
    "errorKey": "UPSTREAM_REQUEST_ERROR",
    "message": "Error while deleting GPR template",
    "data": "Error: gprtemplates.gpr.kubeslice.io \"api-test1\" not found; Reason: NotFound"
  }
}

Create a GPR Template​

Method and URL​

Parameters​

Example Request​

Example Responses​

Get a GPR Template​

Method and URL​

Parameters​

Example Request​

Example Response​

List GPR Templates​

Method and URL​

Parameters​

Example Request​

Example Response​

Update a GPR Template​

Method and URL​

Parameters​

Example Request​

Example Response​

Delete a GPR Template​

Method and URL​

Parameters​

Example Request​

Example Responses​

Create a GPR Template

Method and URL

Parameters

Example Request

Example Responses

Get a GPR Template

Method and URL

Parameters

Example Request

Example Response

List GPR Templates

Method and URL

Parameters

Example Request

Example Response

Update a GPR Template

Method and URL

Parameters

Example Request

Example Response

Delete a GPR Template

Method and URL

Parameters

Example Request

Example Responses