Flower Client Deploy

The deploy action deploys a built Flower client container to participate in federated learning tasks. This action creates and manages the Flower client that communicates with the Flower server.

Overview

The runtime deploys the Flower client container image to a Kubernetes cluster or remote execution environment. The deployed client will automatically join the federated learning federation and participate in training rounds as orchestrated by the Flower server.

Quick example with bare minimum parameters

import digitalhub as dh

# Create Flower client function
f = dh.new_function(
    name="my-flower-client",
    kind="flower-client",
    image="my-registry/my-flower-client:latest",
)

# Deploy the client
run = f.run(action="deploy")

Parameters

Function Parameters

Must be specified when creating the function.

Name	Type	Description
project	str	Project name. Required only when creating from the library; otherwise MUST NOT be set.
name	str	Name that identifies the object. Required.
kind	str	Function kind. Required. MUST BE flower-client
uuid	str	Object ID in UUID4 format.
description	str	Description of the object.
labels	list[str]	List of labels.
embedded	bool	Whether the object should be embedded in the project.
image	str	Docker image name of the built Flower client container. Required for deploy.
base_image	str	Base Docker image used for building (informational only for deploy).
requirements	list[str]	Python package requirements (informational only for deploy).

Task Parameters

Can only be specified when calling function.run().

Name	Type	Description
action	str	Task action. Required. MUST BE deploy
node_selector	list[dict]	Node selector for deployment.
volumes	list[dict]	List of volumes for the deployment.
resources	dict	Resource limits/requests for the deployment.
affinity	dict	Affinity configuration for the deployment.
tolerations	list[dict]	Tolerations for the deployment.
envs	list[dict]	Environment variables for the deployment.
secrets	list[str]	List of secret names for the deployment.
profile	str	Profile template for the deployment.

Run Parameters

Can only be specified when calling function.run().

Name	Type	Description
superlink	str	Flower superlink service endpoint for coordination.
node_config	dict	Node configuration parameters for the Flower client.
root_certificates	str	Content of the root certificate as string.
private_key_secret	str	Name of the secret containing the private key for secure communication.
public_key_secret	str	Name of the secret containing the public key for secure communication.
isolation	str	Isolation mode: process or subprocess.

Entity methods

Run methods

There are no additional run methods for this action.