#Installation

The Data Science Pipelines Operator is delivered as an OLM Operator and installed from the platform OperatorHub.

#Prerequisites

• ACP version: v4.0 or later (validated on v4.3 / Kubernetes 1.34).

• Target cluster architecture linux/amd64 (the operator and its images also ship linux/arm64).

• Operator Lifecycle Manager (OLM) available on the target cluster (provided by ACP).

• The shared Kubeflow Pipelines install (kfp-operator) must NOT be present on the target cluster — DSPO is mutually exclusive with it (see Introduction).

• Argo Workflow CRDs. Each DSPA runs an Argo Workflow controller and the KFP v2 driver/launcher, which require the workflows.argoproj.io CRDs. DSPO does not ship these CRDs in its bundle (to avoid OLM CRD-ownership conflicts on clusters that already run Argo). Install them once at cluster-install time from the operator source tree:

  kubectl apply -k config/argo

  Skip this if the cluster already has workflows.argoproj.io (for example, via an existing Argo install).

• Alauda ServiceMesh v2 (Istio) — only required if you plan to expose the DSPA APIServer through an Istio VirtualService (EXTERNAL_ROUTE_PROVIDER=virtualservice). virtualservices.networking.istio.io must be Established.

#Upload Operator

Download the Data Science Pipelines Operator bundle from the Customer Portal / Marketplace (e.g. data-science-pipelines-operator.ALL.xxxx.tgz), then publish it to the platform repository with the violet command-line tool:

violet push \
  --platform-address=<platform-access-address> \
  --platform-username=<platform-admin> \
  --platform-password=<platform-admin-password> \
  data-science-pipelines-operator.ALL.xxxx.tgz

#Install Operator

In the Administrator view:

1. Click Marketplace / OperatorHub.
2. At the top of the console, from the Cluster dropdown, select the destination cluster.
3. Search for and select Data Science Pipelines Operator, then click Install.
4. Leave Channel unchanged (stable).
5. Check that the Version matches the release you want to install (e.g. v2.15.1).
6. Leave Installation Location unchanged — it defaults to the data-science-pipelines-operator namespace.
7. Choose an Upgrade Strategy (Manual is recommended for production).
8. Click Install.

#External endpoint provider (optional)

The operator exposes each DSPA's APIServer according to its EXTERNAL_ROUTE_PROVIDER setting, configured on the Subscription's spec.config.env:

• none (default) — the operator creates no external endpoint. The APIServer is reachable only in-cluster (ds-pipeline-<name>.<namespace>.svc:8888). Use this when ingress is handled separately.
• virtualservice — the operator creates one Istio VirtualService per DSPA, bound to a shared Istio Gateway (default kubeflow/kubeflow-gateway), routing /pipelines/<namespace>/<name>/... to the DSPA's APIServer and MLMD services. Requires ServiceMesh v2 (see Prerequisites).

To use the virtualservice provider, set the env when creating the Subscription (or edit it afterward):

spec:
  config:
    env:
      - name: EXTERNAL_ROUTE_PROVIDER
        value: virtualservice
      - name: ISTIO_GATEWAY
        value: kubeflow/kubeflow-gateway
      - name: ISTIO_HOSTS
        value: "*"

#Verification

Confirm the Data Science Pipelines Operator tile shows Installed, then verify on the cluster:

# the CSV reports Succeeded
kubectl get csv -n data-science-pipelines-operator | grep data-science-pipelines-operator

# the operator controller-manager pod is Running
kubectl get pods -n data-science-pipelines-operator

# the CRDs are registered
kubectl get crd | grep -E 'datasciencepipelinesapplications|pipelines.kubeflow.org|scheduledworkflows'

The operator watches all namespaces by default, so you can create DataSciencePipelinesApplication resources in any project namespace.