Implementation:SeldonIO Seldon core Seldon Pipeline Load

Field	Value
Implementation Name	Seldon Pipeline Load
Type	External Tool Doc
Overview	Concrete CLI tool for deploying pipelines onto Seldon Core 2.
Related Principle	SeldonIO_Seldon_core_Pipeline_Deployment_Execution
Source	docs-gb/cli/seldon_pipeline_load.md:L1-25
Domains	MLOps, Kubernetes
External Dependencies	seldon CLI, kubectl, Seldon scheduler, Kafka
Knowledge Sources	Repo (https://github.com/SeldonIO/seldon-core), Doc (https://docs.seldon.io/projects/seldon-core/en/v2/)
Last Updated	2026-02-13 00:00 GMT

Description

The seldon pipeline load command submits a Pipeline CRD YAML manifest to the Seldon Core 2 scheduler. The scheduler validates the pipeline topology, provisions Kafka topics for inter-step data flow, and activates the pipeline for inference requests. Alternatively, kubectl apply -f can be used when Seldon is running in Kubernetes mode.

Code Reference

CLI Signature

seldon pipeline load [flags]

CLI Options

Flag	Description	Default
-f, --file-path	Pipeline manifest file (YAML)	(required)
--scheduler-host	Seldon scheduler host	0.0.0.0:9004
--force	Force control plane mode (load pipeline even if validation warnings exist)	false
--authority	Authority (HTTP/2) or virtual host (HTTP/1)	(none)
-v, --verbose	Verbose output	false

Source: docs-gb/cli/seldon_pipeline_load.md:L1-25

I/O Contract

Inputs

• Pipeline CRD YAML: A valid Pipeline manifest file containing apiVersion: mlops.seldon.io/v1alpha1, kind: Pipeline, and the step topology in spec.steps.
• All component models deployed: Every model referenced in spec.steps[].name must be loaded and available on the scheduler.

Outputs

• Pipeline registered with scheduler: The pipeline is submitted and the scheduler begins reconciliation.
• Kafka topics created: Inter-step data flow topics are provisioned for each DAG edge.
• Pipeline version incremented: Each load creates a new version of the pipeline.

Usage Examples

Load a Pipeline via CLI

# Load the tfsimples pipeline
seldon pipeline load -f ./pipelines/tfsimples.yaml

Load with Custom Scheduler Host

# Load pipeline targeting a specific scheduler endpoint
seldon pipeline load -f ./pipelines/tfsimples.yaml --scheduler-host scheduler.seldon.svc:9004

Load via kubectl (Kubernetes Mode)

# Apply the Pipeline CRD directly to Kubernetes
kubectl apply -f ./pipelines/tfsimples.yaml

Full Deployment Workflow

# Step 1: Load component models
seldon model load -f ./models/tfsimple1.yaml
seldon model load -f ./models/tfsimple2.yaml

# Step 2: Wait for models to be available
seldon model status tfsimple1 -w ModelAvailable | jq -M .
seldon model status tfsimple2 -w ModelAvailable | jq -M .

# Step 3: Load the pipeline
seldon pipeline load -f ./pipelines/tfsimples.yaml

# Step 4: Verify pipeline readiness
seldon pipeline status tfsimples -w PipelineReady | jq -M .

Related Pages

• SeldonIO_Seldon_core_Pipeline_Deployment_Execution - implements - Principle describing pipeline deployment and scheduler reconciliation.
• SeldonIO_Seldon_core_Seldon_Pipeline_CRD - input - The Pipeline CRD YAML that is submitted via this command.
• SeldonIO_Seldon_core_Seldon_Model_Load_For_Pipeline - prerequisite - Models must be loaded before the pipeline.
• SeldonIO_Seldon_core_Seldon_Pipeline_Status - next step - Verifying pipeline readiness after loading.
• Environment:SeldonIO_Seldon_core_Kubernetes_Cluster_Environment
• Environment:SeldonIO_Seldon_core_Docker_Compose_Local_Environment
• Environment:SeldonIO_Seldon_core_Kafka_Messaging_Environment
• Heuristic:SeldonIO_Seldon_core_Kafka_Partition_Throughput_Tip