The Kubeflow orchestrator is an orchestrator flavor provided with the ZenML kubeflow integration that uses Kubeflow Pipelines to run your pipelines.

This component is only meant to be used within the context of remote ZenML deployment scenario. Usage with a local ZenML deployment may lead to unexpected behavior!

​
When to use it

You should use the Kubeflow orchestrator if:

  • you’re looking for a proven production-grade orchestrator.

  • you’re looking for a UI in which you can track your pipeline runs.

  • you’re already using Kubernetes or are not afraid of setting up and maintaining a Kubernetes cluster.

  • you’re willing to deploy and maintain Kubeflow Pipelines on your cluster.

​
How to deploy it

The Kubeflow orchestrator supports two different modes: Local and remote. In case you want to run the orchestrator on a local Kubernetes cluster running on your machine, there is no additional infrastructure setup necessary.

If you want to run your pipelines on a remote cluster instead, you’ll need to set up a Kubernetes cluster and deploy Kubeflow Pipelines:

AWS

GCP

Azure

  • Have an existing AWS EKS cluster set up.

  • Make sure you have the AWS CLI set up.

  • Download and install kubectl and configure it to talk to your EKS cluster using the following command:

aws eks --region REGION update-kubeconfig --name CLUSTER_NAME  

gcloud container clusters get-credentials CLUSTER_NAME

  • Install Kubeflow Pipelines onto your cluster.

  • Have an existing AKS cluster set up.

  • Make sure you have the az CLI set up first.

  • Download and install kubectl and it to talk to your AKS cluster using the following command:

az aks get-credentials --resource-group RESOURCE_GROUP --name CLUSTER_NAME
  • Install Kubeflow Pipelines onto your cluster.

Since Kubernetes v1.19, AKS has shifted

to containerd

. However, the workflow controller installed with the Kubeflow installation has Docker set as the default runtime. In order to make your pipelines work, you have to change the value to one of the options

listed here

, preferably k8sapi.

This change has to be made by editing the containerRuntimeExecutor property of the ConfigMap corresponding to the workflow controller. Run the following commands to first know what config map to change and then to edit it to reflect your new value.

kubectl get configmap -n kubeflow
kubectl edit configmap CONFIGMAP_NAME -n kubeflow
# This opens up an editor that can be used to make the change.

If one or more of the deployments are not in the Running state, try increasing the number of nodes in your cluster.

If you’re installing Kubeflow Pipelines manually, make sure the Kubernetes service is called exactly ml-pipeline. This is a requirement for ZenML to connect to your Kubeflow Pipelines deployment.

​
How to use it

To use the Kubeflow orchestrator, we need:

  • The ZenML kubeflow integration installed. If you haven’t done so, run
zenml integration install kubeflow

Local

Remote

When using the Kubeflow orchestrator locally, you’ll additionally need:

The local Kubeflow Pipelines deployment requires more than 2 GB of RAM, so if you’re using Docker Desktop make sure to update the resource limits in the preferences.

We can then register the orchestrator and use it in our active stack:

zenml orchestrator register  \
    --flavor=kubeflow

# Register and activate a stack with the new orchestrator
zenml stack register  -o  ... --set

When using the Kubeflow orchestrator with a remote cluster, you’ll additionally need:

  • A remote ZenML server deployed to the cloud. See the deployment guide for more information.

  • Kubeflow pipelines deployed on a remote cluster. See the deployment section for more information.

  • The name of your Kubernetes context which points to your remote cluster. Run kubectl config get-contexts to see a list of available contexts.

  • A remote artifact store as part of your stack.

  • A remote container registry as part of your stack.

We can then register the orchestrator and use it in our active stack:

zenml orchestrator register  \
    --flavor=kubeflow \
    --kubernetes_context=

# Add the orchestrator to the active stack
zenml stack update -o

ZenML will build a Docker image called <CONTAINER_REGISTRY_URI>/zenml:<PIPELINE_NAME> which includes your code and use it to run your pipeline steps in Kubeflow. Check out this page if you want to learn more about how ZenML builds these images and how you can customize them.

Once the orchestrator is part of the active stack, we need to run zenml stack up before running any pipelines. This command

  • forwards a port, so you can view the Kubeflow UI in your browser.

  • (in the local case) uses K3D to provision a Kubernetes cluster on your machine and deploys Kubeflow Pipelines on it.

You can now run any ZenML pipeline using the Kubeflow orchestrator:

python file_that_runs_a_zenml_pipeline.py

​
Additional configuration

For additional configuration of the Kubeflow orchestrator, you can pass KubeflowOrchestratorSettings which allows you to configure (among others) the following attributes:

  • client_args: Arguments to pass when initializing the KFP client.

  • user_namespace: The user namespace to use when creating experiments and runs.

  • pod_settings: Node selectors, affinity and tolerations to apply to the Kubernetes Pods running your pipline. These can be either specified using the Kubernetes model objects or as dictionaries.

from zenml.integrations.kubeflow.flavors.kubeflow_orchestrator_flavor import KubeflowOrchestratorSettings
from kubernetes.client.models import V1Toleration

kubeflow_settings = KubeflowOrchestratorSettings(
    client_args={},
    user_namespace="my_namespace",
    pod_settings={
        "affinity": {
            "nodeAffinity": {
                "requiredDuringSchedulingIgnoredDuringExecution": {
                    "nodeSelectorTerms": [
                        {
                            "matchExpressions": [
                                {
                                    "key": "node.kubernetes.io/name",
                                    "operator": "In",
                                    "values": ["my_powerful_node_group"],
                                }
                            ]
                        }
                    ]
                }
            }
        },
        "tolerations": [
            V1Toleration(
                key="node.kubernetes.io/name",
                operator="Equal",
                value="",
                effect="NoSchedule"
            )
        ]
    }
)

@pipeline(
    settings={
        "orchestrator.kubeflow": kubeflow_settings
    }
)
    ...

Check out the API docs for a full list of available attributes and this docs page for more information on how to specify settings.

​
Enabling CUDA for GPU-backed hardware

Note that if you wish to use this orchestrator to run steps on a GPU, you will need to follow the instructions on this page to ensure that it works. It requires adding some extra settings customization and is essential to enable CUDA for the GPU to give its full acceleration.

​
Important Note for Multi-Tenancy Deployments

Kubeflow has a notion of multi-tenancy built into its deployment. Kubeflow’s multi-user isolation simplifies user operations because each user only views and edited\s the Kubeflow components and model artifacts defined in their configuration.

Using the ZenML Kubeflow orchestrator on a multi-tenant deployment without any settings will result in the following error:

HTTP response body: {"error":"Invalid input error: Invalid resource references for experiment. ListExperiment requires filtering by namespace.","code":3,"message":"Invalid input error: Invalid resource references for experiment. ListExperiment requires filtering by 
namespace.","details":[{"@type":"type.googleapis.com/api.Error","error_message":"Invalid resource references for experiment. ListExperiment requires filtering by namespace.","error_details":"Invalid input error: Invalid resource references for experiment. ListExperiment requires filtering by namespace."}]}

In order to get it to work, we need to leverage the KubeflowOrchestratorSettings referenced above. By setting the namespace option, and by passing in the right authentication credentials to the Kubeflow Pipelines Client, we can make it work.

First, when registering your kubeflow orchestrator, please make sure to include the kubeflow_hostname parameter. The kubeflow_hostname must end with the /pipeline post-fix.

zenml orchestrator register  \
    --flavor=kubeflow \
    --kubernetes_context= \  
    --kubeflow_hostname= # e.g. https://mykubeflow.example.com/pipeline

Then, ensure that you use the pass the right settings before triggering a pipeline run. The following snippet will prove useful:

import requests

from zenml.client import Client
from zenml.integrations.kubeflow.flavors.kubeflow_orchestrator_flavor import (
    KubeflowOrchestratorSettings,
)

NAMESPACE = "namespace_name"  # This is the user namespace for the profile you want to use
USERNAME = "admin"  # This is the username for the profile you want to use
PASSWORD = "abc123"  # This is the password for the profile you want to use

# Use client_username and client_password and ZenML will automatically fetch a session cookie
kubeflow_settings = KubeflowOrchestratorSettings(
    client_username=USERNAME,
    client_password=PASSWORD,
    user_namespace=NAMESPACE
)

# You can also pass the cookie in `client_args` directly
# kubeflow_settings = KubeflowOrchestratorSettings(
#     client_args={"cookies": session_cookie}, user_namespace=NAMESPACE
# )

@pipeline(
    settings={
        "orchestrator.kubeflow": kubeflow_settings
    }
):
    ...

if "__name__" == "__main__":
  # Run the pipeline

Note that the above is also currently not tested on all Kubeflow versions, so there might be further bugs with older Kubeflow versions. In this case, please reach out to us on Slack.

​
Using secrets in settings

The above example encoded the username and password in plain-text as settings. You can also set them as secrets.

zenml secrets-manager secret register kubeflow_secret \
    --username=admin \
    --password=abc123

And then you can use them in code:

# Use client_username and client_password and ZenML will automatically fetch a session cookie
kubeflow_settings = KubeflowOrchestratorSettings(
    client_username="{{kubeflow_secret.username}}",  # secret reference
    client_password="{{kubeflow_secret.password}}",  # secret reference
    user_namespace="namespace_name"
)

See full documentation of using secrets within ZenML here.

A concrete example of using the Kubeflow orchestrator can be found here.

For more information and a full list of configurable attributes of the Kubeflow orchestrator, check out the API Docs.