functions

Creates, updates, deletes, gets or lists a functions resource.

Overview

Name	functions
Type	Resource
Id	nvidia.nvcf_functions.functions

Fields

The following fields are returned by SELECT queries:

list

Name	Datatype	Description
id	string (uuid)	Unique function id
name	string (UTF-8)	Function name
nca_id	string (UTF-8)	NVIDIA Cloud Account Id (wire: ncaId)
version_id	string (uuid)	Unique function version id (wire: versionId)
helm_chart_service_name	string (UTF-8)	Helm Chart Service Name specified only when helmChart property is specified (wire: helmChartServiceName)
active_instances	array	List of active instances for this function. (wire: activeInstances)
api_body_format	string (UTF-8)	Invocation request body format (PREDICT_V2, CUSTOM) (wire: apiBodyFormat)
container_args	string (UTF-8)	Args used to launch the container (wire: containerArgs)
container_environment	array	Environment settings used to launch the container (wire: containerEnvironment)
container_image	string (uri)	Optional custom container (wire: containerImage)
created_at	string (date_time)	Function creation timestamp (wire: createdAt)
description	string (UTF-8)	Function/version description
function_type	string (UTF-8)	Used to indicate a STREAMING function. Defaults to DEFAULT. (DEFAULT, STREAMING, LLM) (wire: functionType)
health	object	Function health configuration
health_uri	string (uri)	Health endpoint for the container or helmChart. Deprecated, use health.uri instead. (wire: healthUri)
helm_chart	string (uri)	Optional Helm Chart (wire: helmChart)
inference_port	integer (int32)	Optional port number where the inference listener is running - defaults to 8000 for Triton (wire: inferencePort)
inference_url	string (uri)	Entrypoint for invoking the container to process requests (wire: inferenceUrl)
llm_invocation_config	object	Optional function-level LLM invocation configuration. (wire: llmInvocationConfig)
models	array	Optional list of models
owned_by_different_account	boolean	Indicates whether the function is owned by another account. If the account that is being used to lookup functions happens to be authorized to invoke/list this function which is owned by a different account, then this field is set to true and ncaId will contain the id of the account that owns the function. Otherwise, this field is not set as it defaults to false. (wire: ownedByDifferentAccount)
rate_limit	object	Optional rate limit policy (wire: rateLimit)
resources	array	Optional set of resources.
secrets	array	Optional secret names
status	string (UTF-8)	Function status (ACTIVE, DEPLOYING, ERROR, INACTIVE, DELETED, DEGRADED, DEGRADING)
tags	array	Optional set of tags. Maximum allowed number of tags per function is 64. Maximum length of each tag is 128 chars.
telemetries	object	Optional telemetry configuration for the function

Methods

The following methods are available for this resource:

Name	Accessible by	Required Params	Optional Params	Description
list	select		visibility	Lists all the functions associated with the authenticated NVIDIA Cloud Account. Requires a bearer token with 'list_functions' or 'list_functions_details' scope in the HTTP Authorization header.
create	insert	inference_url, name		Creates a new function within the authenticated NVIDIA Cloud Account. Requires a bearer token with 'register_function' scope in the HTTP Authorization header.

Parameters

Parameters can be passed in the WHERE clause of a query. Check the Methods section to see which parameters are required or optional for each operation.

Name	Datatype	Description
visibility	array	Query param 'visibility' indicates the kind of functions to be included in the response.

SELECT examples

list

Lists all the functions associated with the authenticated NVIDIA Cloud Account. Requires a bearer token with 'list_functions' or 'list_functions_details' scope in the HTTP Authorization header.

SELECT
id,
name,
nca_id,
version_id,
helm_chart_service_name,
active_instances,
api_body_format,
container_args,
container_environment,
container_image,
created_at,
description,
function_type,
health,
health_uri,
helm_chart,
inference_port,
inference_url,
llm_invocation_config,
models,
owned_by_different_account,
rate_limit,
resources,
secrets,
status,
tags,
telemetries
FROM nvidia.nvcf_functions.functions
WHERE visibility = '{{ visibility }}'
;

INSERT examples

create

Creates a new function within the authenticated NVIDIA Cloud Account. Requires a bearer token with 'register_function' scope in the HTTP Authorization header.

INSERT INTO nvidia.nvcf_functions.functions (
api_body_format,
container_args,
container_environment,
container_image,
description,
function_type,
health,
health_uri,
helm_chart,
helm_chart_service_name,
inference_port,
inference_url,
llm_invocation_config,
models,
name,
rate_limit,
resources,
secrets,
tags,
telemetries
)
SELECT 
'{{ api_body_format }}',
'{{ container_args }}',
'{{ container_environment }}',
'{{ container_image }}',
'{{ description }}',
'{{ function_type }}',
'{{ health }}',
'{{ health_uri }}',
'{{ helm_chart }}',
'{{ helm_chart_service_name }}',
{{ inference_port }},
'{{ inference_url }}' /* required */,
'{{ llm_invocation_config }}',
'{{ models }}',
'{{ name }}' /* required */,
'{{ rate_limit }}',
'{{ resources }}',
'{{ secrets }}',
'{{ tags }}',
'{{ telemetries }}'
RETURNING
function
;

Manifest

# Description fields are for documentation purposes
- name: functions
props:
  - name: api_body_format
    value: "{{ api_body_format }}"
    description: |
      Invocation request body format
    valid_values: ['PREDICT_V2', 'CUSTOM']
  - name: container_args
    value: "{{ container_args }}"
    description: |
      Args to be passed when launching the container
  - name: container_environment
    description: |
      Environment settings for launching the container
    value:
      - key: "{{ key }}"
        value: "{{ value }}"
  - name: container_image
    value: "{{ container_image }}"
    description: |
      Optional custom container image
  - name: description
    value: "{{ description }}"
    description: |
      Optional function/version description
  - name: function_type
    value: "{{ function_type }}"
    description: |
      Optional function type, used to indicate a STREAMING function. Defaults to DEFAULT.
    valid_values: ['DEFAULT', 'STREAMING', 'LLM']
  - name: health
    description: |
      Function health
    value:
      expectedStatusCode: {{ expectedStatusCode }}
      port: {{ port }}
      protocol: "{{ protocol }}"
      timeout: "{{ timeout }}"
      uri: "{{ uri }}"
  - name: health_uri
    value: "{{ health_uri }}"
    description: |
      Health endpoint for the container or the helmChart
  - name: helm_chart
    value: "{{ helm_chart }}"
    description: |
      Optional Helm Chart
  - name: helm_chart_service_name
    value: "{{ helm_chart_service_name }}"
    description: |
      Helm Chart Service Name is required when helmChart property is specified
  - name: inference_port
    value: {{ inference_port }}
    description: |
      Optional port number where the inference listener is running. Defaults to 8000
      for Triton.
  - name: inference_url
    value: "{{ inference_url }}"
    description: |
      Entrypoint for invoking the container to process a request
  - name: llm_invocation_config
    description: |
      Optional function-level LLM invocation configuration. Only valid for LLM-type functions.
    value:
      priority:
        defaultPriority: {{ defaultPriority }}
        perAccountPriority: "{{ perAccountPriority }}"
  - name: models
    description: |
      Optional list of models
    value:
      - llmConfig:
          routingMethod: "{{ routingMethod }}"
          tokenRateLimit: "{{ tokenRateLimit }}"
          tokenizer: "{{ tokenizer }}"
          uris:
            - "{{ uris }}"
        name: "{{ name }}"
        uri: "{{ uri }}"
        version: "{{ version }}"
  - name: name
    value: "{{ name }}"
    description: |
      Function name must start with lowercase/uppercase/digit and can only contain lowercase, uppercase, digit, hyphen, and underscore characters
  - name: rate_limit
    description: |
      Optional rate limit config
    value:
      exemptedNcaIds:
        - "{{ exemptedNcaIds }}"
      perNcaIdRate: "{{ perNcaIdRate }}"
      perUserRate: "{{ perUserRate }}"
      rateLimit: "{{ rateLimit }}"
      syncCheck: {{ syncCheck }}
  - name: resources
    description: |
      Optional set of resources
    value:
      - name: "{{ name }}"
        uri: "{{ uri }}"
        version: "{{ version }}"
  - name: secrets
    description: |
      Optional secrets
    value:
      - name: "{{ name }}"
        value: "{{ value }}"
  - name: tags
    value:
      - "{{ tags }}"
    description: |
      Optional set of tags - could be empty. Provided by user
  - name: telemetries
    description: |
      Optional telemetry configuration for logs, metrics, and traces.
    value:
      logsTelemetryId: "{{ logsTelemetryId }}"
      metricsTelemetryId: "{{ metricsTelemetryId }}"
      tracesTelemetryId: "{{ tracesTelemetryId }}"