Evaluating engineering tools? Get the comparison in Google Sheets

Installation and Setup

helm install seldon-core seldon-core/seldon-core-operator --namespace seldon-system --create-namespace
Install Seldon Core using Helm

kubectl create namespace seldon-system
Create namespace for Seldon Core

kubectl apply -f https://raw.githubusercontent.com/SeldonIO/seldon-core/master/helm-charts/seldon-core-operator/templates/seldon-core.yaml
Deploy Seldon Core operator without Helm

Deployment Management

kubectl apply -f seldon-deployment.yaml
Deploy a Seldon model

kubectl get svc -n
Get services to find model endpoint

kubectl get seldondeployments -n
List Seldon deployments

kubectl describe seldondeployment -n
Describe a Seldon deployment

kubectl delete seldondeployment -n
Delete a Seldon deployment

Monitoring

kubectl port-forward svc/seldon-core-analytics-prometheus -n seldon-system 9090:9090
Access Prometheus for metrics

kubectl port-forward svc/seldon-core-analytics-grafana -n seldon-system 3000:80
Access Grafana dashboards

kubectl logs deploy/seldon-controller-manager -n seldon-system
Check operator logs

Testing and Inference

curl -X POST -H 'Content-Type: application/json' -d '{"data": {"ndarray": [[...data...]]}}' http:///api/v1.0/predictions
Send prediction request

seldon-core-api-tester --host --endpoint --namespace
Test API with tester tool

seldon-core-microservice --api-type REST
Run a model locally for testing

Troubleshooting

kubectl get pods -n | grep seldon
List all Seldon pods

kubectl logs -n
View logs for a specific pod

kubectl describe pod -n
Get detailed pod information

kubectl get events -n
Get events for troubleshooting