create-ml-serving-app¶

Description¶

Deploys Cloudera AI Inference Service into an existing Kubernetes cluster.

Synopsis¶

  create-ml-serving-app
--app-name <value>
--environment-crn <value>
[--cluster-crn <value>]
[--provision-k8s-request <value>]
[--use-public-load-balancer | --no-use-public-load-balancer]
[--skip-validation | --no-skip-validation]
[--load-balancer-ip-whitelists <value>]
[--subnets-for-load-balancers <value>]
[--static-subdomain <value>]
[--ozone-s3-creds <value>]
[--enable-fine-grained-access-control | --no-enable-fine-grained-access-control]
[--cli-input-json <value>]
[--generate-cli-skeleton]

Options¶

--app-name (string)

The name of the Cloudera AI Inference Service to be created.

--environment-crn (string)

The environment CRN.

--cluster-crn (string)

The cluster CRN of an existing cluster where Cloudera AI Inference Service instance will be deployed.

--provision-k8s-request (object)

Request object for provisioning Cloudera AI Inference Service instance groups.

environmentCrn -> (string)

The CRN of the environment for the cluster to create.

instanceGroups -> (array)

The instance groups.

item -> (object)

Contains the necessary info for an instance group.

instanceType -> (string)

The cloud provider instance type for the node instance.

instanceTier -> (string)

The tier of the instance i.e. on-demand/spot.

instanceCount -> (integer)

The initial number of instance node.

name -> (string)

The unique name for the instance or resource group of the workbench.

ingressRules -> (array)

The networking rules for the ingress.

item -> (string)

rootVolume -> (object)

Configuration for instance root device volume.

size -> (integer)

The volume size in GB.

autoscaling -> (object)

Configuration for instance auto scaling.

minInstances -> (integer)

The minimum number of instance for auto scaling.

maxInstances -> (integer)

The maximum number of instance for auto scaling.

enabled -> (boolean)

The boolean flag to enable the auto scaling.

accelerator -> (object)

Accelerator, i.e. GPU configurations.

type -> (string)

The GPU type.

count -> (integer)

The GPU count.

tags -> (array)

Tags to add to the cloud provider resources created. This is in addition to any tags added by Cloudera.

item -> (object)

Contains a single tag entry that will be configured on cloud resources associated with a workbench.

key -> (string)

The name for the tag.

value -> (string)

The value for the tag.

network -> (object)

Contains the information about overlay network.

plugin -> (string)

The plugin specifies specific cni vendor, ex: calico, weave etc.

topology -> (object)

Contains the information about topology.

subnets -> (array)

The options for subnets.

item -> (string)

gcp -> (object)

Contains the information about gcp.

secondaryPodRangeName -> (string)

The secondary pod range name for GCP.

JSON Syntax:

{
  "environmentCrn": "string",
  "instanceGroups": [
    {
      "instanceType": "string",
      "instanceTier": "string",
      "instanceCount": integer,
      "name": "string",
      "ingressRules": ["string", ...],
      "rootVolume": {
        "size": integer
      },
      "autoscaling": {
        "minInstances": integer,
        "maxInstances": integer,
        "enabled": true|false
      },
      "accelerator": {
        "type": "string",
        "count": integer
      }
    }
    ...
  ],
  "tags": [
    {
      "key": "string",
      "value": "string"
    }
    ...
  ],
  "network": {
    "plugin": "string",
    "topology": {
      "subnets": ["string", ...]
    },
    "gcp": {
      "secondaryPodRangeName": "string"
    }
  }
}

--use-public-load-balancer | --no-use-public-load-balancer (boolean)

The boolean flag to request a public load balancer. By default, a private load balancer is used.

--skip-validation | --no-skip-validation (boolean)

Skip pre-flight validations if requested.

--load-balancer-ip-whitelists (array)

The whitelist of IPs for load balancer.

Syntax:

"string" "string" ...

--subnets-for-load-balancers (array)

The list of subnets to be used for the load balancer.

Syntax:

"string" "string" ...

--static-subdomain (string)

The static subdomain to be used for the Cloudera AI Inference Service instance.

--ozone-s3-creds (object)

Request object for the MlServingOzoneCreds method.

s3AccessID -> (string)

The Ozone S3 access key ID name.

s3SecretAccessKey -> (string)

The Ozone S3 secret access key name.

s3EndPointUrl -> (string)

The Ozone S3 end point.

s3CaCertBundle -> (string)

The Ozone S3 CA bundle.

Shorthand Syntax:

s3AccessID=string,s3SecretAccessKey=string,s3EndPointUrl=string,s3CaCertBundle=string

JSON Syntax:

{
  "s3AccessID": "string",
  "s3SecretAccessKey": "string",
  "s3EndPointUrl": "string",
  "s3CaCertBundle": "string"
}

--enable-fine-grained-access-control | --no-enable-fine-grained-access-control (boolean)

The boolean flag to enable fine-grained access control. By default, this flag is disabled.

--cli-input-json (string)

Performs service operation based on the JSON string provided. The JSON string follows the format provided by --generate-cli-skeleton. If other arguments are provided on the command line, the CLI values will override the JSON-provided values.

--generate-cli-skeleton (boolean)

Prints a sample input JSON to standard output. Note the specified operation is not run if this argument is specified. The sample input can be used as an argument for --cli-input-json.

Output¶

appCrn -> (string)

The CRN of the Cloudera AI Inference Service created.

Form Factors¶

public, private