Implementation:Kserve Kserve Component ISVC Deployment

Knowledge Sources	KServe KServe InferenceGraph
Domains	MLOps, Pipeline, Model_Serving
Last Updated	2026-02-13 00:00 GMT

Overview

Concrete YAML pattern for deploying multiple InferenceService resources as components for InferenceGraph pipelines.

Description

Graph pipeline components are standard InferenceService resources deployed using kubectl apply. The sample YAML files demonstrate deploying multiple models (sklearn, xgboost) and custom containers as individual services. These services are later referenced by the InferenceGraph spec using serviceName.

Usage

Deploy all component ISVCs and verify they are ready before creating the InferenceGraph that references them.

Code Reference

Source Location

• Repository: kserve
• File: docs/samples/graph/sequence.yaml, Lines 1-18 (sklearn + xgboost ISVCs)
• File: docs/samples/graph/switch.yaml, Lines 1-41 (custom container ISVCs)

Signature

# Multiple component ISVCs in one file
apiVersion: serving.kserve.io/v1beta1
kind: InferenceService
metadata:
  name: sklearn-iris
spec:
  predictor:
    sklearn:
      storageUri: "gs://kfserving-examples/models/sklearn/1.0/model"
---
apiVersion: serving.kserve.io/v1beta1
kind: InferenceService
metadata:
  name: xgboost-iris
spec:
  predictor:
    xgboost:
      storageUri: "gs://kfserving-examples/models/xgboost/iris"

Import

kubectl apply -f sequence.yaml

I/O Contract

Inputs

Name	Type	Required	Description
Model artifacts	GCS/S3/container	Yes	Model files or container images for each component
Component names	string	Yes	Unique names used as serviceName in graph spec

Outputs

Name	Type	Description
InferenceService endpoints	URL	Running prediction endpoints for each component
Service names	string	Names used to reference components in InferenceGraph

Usage Examples

Deploy Sequence Components

# Deploy sklearn and xgboost components
kubectl apply -f - <<EOF
apiVersion: serving.kserve.io/v1beta1
kind: InferenceService
metadata:
  name: sklearn-iris
spec:
  predictor:
    sklearn:
      storageUri: "gs://kfserving-examples/models/sklearn/1.0/model"
---
apiVersion: serving.kserve.io/v1beta1
kind: InferenceService
metadata:
  name: xgboost-iris
spec:
  predictor:
    xgboost:
      storageUri: "gs://kfserving-examples/models/xgboost/iris"
EOF

# Wait for all components
kubectl wait inferenceservice sklearn-iris --for=condition=Ready
kubectl wait inferenceservice xgboost-iris --for=condition=Ready

Related Pages

Implements Principle

• Principle:Kserve_Kserve_Graph_Component_Deployment