Google Kubernetes Engine (GKE) requirements
Google Kubernetes Engine (GKE)¶
This section outlines the supplementary configuration required for the Google Kubernetes Engine (GKE) to successfully run the DataRobot platform.
Note
You must fulfill the general Kubernetes requirements before proceeding.
The services below are assumed in a GKE deployment because the target environment is a Google Cloud project, but they are not required. This page does not cover setup of clusters without these services.
GKE service integration¶
| Amazon service | Reference |
|---|---|
| GKE | See Supported Kubernetes versions. |
| Cloud Storage | Object storage and Image Build Service cache and logging (Google Documentation) |
| Artifact Registry | Primary registry and Image Build Service repository (Google Documentation) |
| OIDC | Workload Identity Federation with GKE |
| Service account key (optional) | Service account key |
Google Artifact Registry (GAR)¶
You can install the DataRobot platform by pulling the application images from your private Google Artifact Registry (GAR). Alternatively, you can use the Docker Hub registry.
Note
Private GAR is required to use custom models functionality in DataRobot.
See the Artifact Registry documentation Google Cloud documentation for details.
Network policy engine¶
Custom model and custom task features in DataRobot benefit from a network policy engine to enforce Kubernetes Network Policies. It is recommended, but not required, to ensure the security of custom model pods.
See the Control communication between Pods and Services using network policies Google Cloud documentation for details.