Ray

KubeRay is an open-source toolkit designed to facilitate the execution of Ray applications on Kubernetes. It offers a range of tools that enhance the operational aspects of running and overseeing Ray on Kubernetes.

Key components include:

• Ray Operator
• Backend services for cluster resource creation and deletion
• Kubectl plugin/CLI for CRD object management
• Seamless integration of Jobs and Serving functionality with Clusters

Install the plugin

To install the Ray plugin, run the following command:

$ pip install flytekitplugins-ray

To enable the plugin in the backend, refer to the instructions provided in the Configuyre Kubernetes Pluginsep guide.

Implementation details

Submit a Ray job to existing cluster

import ray
from flytekit import task
from flytekitplugins.ray import RayJobConfig
@ray.remote
def f(x):
    return x * x
@task(
    task_config=RayJobConfig(
        address=<RAY_CLUSTER_ADDRESS>
        runtime_env={"pip": ["numpy", "pandas"]}
    )
)
def ray_task() -> typing.List[int]:
    futures = [f.remote(i) for i in range(5)]
    return ray.get(futures)

Create a Ray cluster managed by Flyte and run a Ray Job on the cluster

import ray
from flytekit import task
from flytekitplugins.ray import RayJobConfig, WorkerNodeConfig, HeadNodeConfig
@task(task_config=RayJobConfig(worker_node_config=[WorkerNodeConfig(group_name="test-group", replicas=10)]))
def ray_task() -> typing.List[int]:
    futures = [f.remote(i) for i in range(5)]
    return ray.get(futures)

Run the example on the Flyte cluster

To run the provided example on the Flyte cluster, use the following command:

$ pyflyte run --remote ray_example.py \
          ray_workflow --n 10