HUhuggingface
Endpoints

Scale endpoint to zero [WRITE]

Changed on

Scale the endpoint down to zero replicas and mark it as Stopped. The endpoint's spec must have scale-to-zero enabled (scaling.scaleToZeroTimeout set); otherwise the call returns 400 "workload cannot be scaled to zero". Not valid when the workload is in the Deleting stage. Use POST /resume to wake the endpoint back up.

post/v2/endpoint/{namespace}/{name}/scale-to-zero

Request

  • Base URL
    https://api.endpoints.huggingface.cloud (the document declares no server; this is the origin it was published from)
  • URL
    https://api.endpoints.huggingface.cloud/v2/endpoint/{namespace}/{name}/scale-to-zero
  • Auth
    HTTP bearer

Path parameters

namespacestring required

User or organization name

namestring required

Endpoint name

Response

Endpoint scaled to zero successfully

healthRoutestring required

Endpoint health route

namestring required

Endpoint name. Must only contains lowercase alphanumeric characters or '-' and have a length of 32 characters maximum

tagsstring[] required
type'public' | 'authenticated' | 'private' required

Endpoint visibility and authentication mode

Example response

{
  "compute": {
    "instanceSize": "x1",
    "instanceType": "intel-spr",
    "scaling": {
      "maxReplica": 4,
      "measure": {
        "hardwareUsage": 80
      },
      "minReplica": 1,
      "scaleToZeroTimeout": 15,
      "schedule": [
        {
          "at": "07:30",
          "maxReplica": 4,
          "minReplica": 1,
          "scaleToZeroTimeout": 15
        }
      ]
    }
  },
  "model": {
    "env": {
      "MAX_CONCURRENT_REQUESTS": "128"
    },
    "image": {
      "tgi": {
        "healthRoute": "/health",
        "port": 80,
        "url": "https://host/image:tag",
        "maxBatchPrefillTokens": 2048,
        "maxBatchTotalTokens": 2048,
        "maxInputLength": 1024,
        "maxTotalTokens": 1512,
        "disableCustomKernels": true
      }
    },
    "repository": "meta-llama/Llama-2-7b-hf",
    "revision": "main",
    "secrets": {
      "HF_TOKEN": "hf_xxx"
    }
  },
  "name": "my-endpoint",
  "network": {
    "cidrList": [
      "10.0.0.0/8"
    ]
  },
  "privateService": {
    "accountId": "123456789012"
  },
  "provider": {
    "region": "us-east-1",
    "vendor": "aws"
  },
  "route": {
    "domain": "api.example.com",
    "path": "/v1/models/gpt2"
  },
  "status": {
    "readyReplica": 2,
    "targetReplica": 4,
    "url": "https://68a0f3b5c6ee5c9dcc0e0a01.endpoints.huggingface.cloud"
  }
}

Changes

  • b42f445e0429Revision changes
    • breaking

      added subschema #2 subschema #3 subschema #4 subschema #5 subschema #6 subschema #7 subschema #8 subschema #10 to the / response property oneOf list for the response status

    • info

      added the optional property // to the response with the status

    • info

      removed subschema #2 subschema #3 subschema #4 subschema #5 subschema #6 subschema #7 subschema #8 subschema #10 from the / response property oneOf list for the response status

  • 6fca5be0e3e9Revision changes
    • breaking

      added subschema #3 to the ////// response property oneOf list for the response status