Google Kubernetes Engine (GKE)¶
This section describes the GKE cluster requirements for running the DataRobot application. It assumes that you have provisioned sufficient hardware resources to run large computational workloads.
Terraform infrastructure module¶
To eliminate the need for manual operations and allow for simplified configuration, DataRobot provides a dedicated Terraform module for GCP infrastructure.
This module automates the setup of critical components, including:
- Network: An Google VPC with the necessary subnetting and a Cloud NAT gateway.
- Kubernetes: Provisioning an Google Kubernetes Engine (GKE) cluster.
- Storage: An Google Cloud Storage (GCS) bucket for the DataRobot application.
- Container Registry: An Google Artifact Registry (GAR) to host custom images built and used by DataRobot services.
- Databases:
- Cloud SQL for PostgreSQL
- Memorystore Redis
- MongoDB Atlas cluster (for use by the DataRobot application)
- Networking Add-ons (via Helm Charts):
ingress-nginxfor managing external access to the Kubernetes cluster.cert-managerfor automated certificate management (e.g., Let's Encrypt).external-dnsfor automatically creating DNS records for Kubernetes services.
This simplifies and standardizes the infrastructure layer required before installing the DataRobot platform via Helm charts.
Platform-specific requirements¶
In addition to the generic infrastructure requirements, you need the following platform-specific requirements to install DataRobot in the Google cloud:
- Provisioner host requirements.
- Google Kubernetes Engine (GKE) requirements.
- OIDC provider.
- Google Cloud Storage.
Preparation¶
You can install the DataRobot platform by pulling the application images directly from Docker Hub.
If you use a private Google Artifact Registry (GAR), you must first pull the DataRobot container images and push them to your private OCI registry before starting the installation.
設定¶
Prepare your values_dr.yaml file with all the necessary configurations for your environment: