ML Service Chart

The ML Service manages machine learning model training, serving, and lifecycle with Ray cluster integration and MLflow experiment tracking.

Chart Configuration

ml-service:
  enabled: true
  replicaCount: 1
 
  resources:
    requests:
      cpu: "500m"
      memory: "1Gi"
    limits:
      cpu: "4"
      memory: "8Gi"
 
  config:
    ray:
      headAddress: "ray-head:6379"
      dashboardUrl: "http://ray-head:8265"
    mlflow:
      trackingUrl: "http://mlflow:5000"
    serve:
      enabled: true
      port: 8000

ML Infrastructure Connections

Component	Purpose	Port
Ray Head	Distributed training orchestration	6379, 8265
MLflow	Experiment tracking, model registry	5000
MinIO/S3	Model artifact storage	9000
Feast	Feature store	6566

Pipeline Service BI Service