Skip to content

Server Setup

Introduction

FeatureByte is designed to be deployed in a kubernetes cluster. This guide provides step-by-step walkthrough on how to set up and configure the FeatureByte server.

Prerequisites

Server Location

Workflow Diagram

FeatureByte should be placed in a secure network zone, ideally within the same cloud provider as your data warehouse. This minimizes latency and enhances security by reducing data exposure.

FeatureByte interacts with your provided data warehouse and a blob storage provider. Ensure that the server has network routes to these services. For example, if your organisation uses Azure OpenAI, ensure that the server can access the Azure OpenAI endpoints, likewise if you are using the public OpenAI endpoints, ensure that your cluster has internet access.

FeatureByte Images

FeatureByte platform is distributed as a set of Docker images. For uses on AWS, we are able to share the images directly by granting permissions on EKS cluster node IAM role. For other cloud providers, we will ship you the images and you will need to upload them to your private container registry.

Node Group Sizing

Due to the unpredictable nature of ML workloads. we recommend starting with a node group of at least 3 nodes with 16 vCPU and 64GB of RAM. Depending on the datasize and complexity of the features you are creating, you may need to scale the node group further, both horizontally and vertically.

Ingress

FeatureByte does not come with a specific ingress requirement. You can use any ingress controller of your choice, please involve us if you are unsure about the configuration.

Post-Installation Configuration

For featurebyte to perform optimally, there are a few configurations that need to be done after the installation.

Settings

Go to the > Admin > Settings page. Here you can configure the following settings:

  1. SSO Settings: If you are using SSO, configure the SSO settings here. You can choose from Google, Microsoft, Okta and Sailpoint as your SSO provider. You can also set the default RBAC role for new users here.
  2. GenAI Settings: Configure the GenAI settings here. You can choose from OpenAI and Azure OpenAI as your GenAI provider. You will need to provide the API key and the model you want to use.
  3. License Settings: Please input the license key provided to you by FeatureByte.
  4. Data Ware Connection: Configure the connection to your data warehouse here. You will need to provide the connection details and a credential that has access to the data warehouse.