Creating Kyvos resources for AWS automated deployment with Databricks

Use Graviton Instances

Select the checkbox to deploy the Kyvos Services (like BI Server, Query Engine, and Kyvos Manager) on Graviton instances.

EC2 instance(s) for BI Server

• Number of Node(s)

• Instance Type

• Installation Disk

EC2 instance(s) for Query Engine

• Number of Node(s)

• Instance Type

• Installation Disk

• EBS Volume

NOTE: Installation disk size must be 100 GB, and the cuboid replication disk size must be in the multiplier of 50 or 100.  

Web Portal High Availability 

Select to enable web portal high availability, and specify the following:

• Number of Node(s) 

• Instance Type 

• Installation Disk 

• Subnet for Load Balancer  (Possible value  list should be Public or Private)

EC2 instance for Kyvos Manager

The checkbox is selected by default for the automated deployment and an EC2 instance is launched for the Kyvos Manager.

• Provide the Instance Type and Installation Disk details.
  When using a multi-tenant environment, you can use the Deploy as Tenant checkbox to define a tenant.

• Select one of the following repository types from the  Repository  list.

  • Bundled Postgres: To select Postgres as a repository for Kyvos Manager. 

  • Create New RDS: To create a new external repository for Kyvos Manager.

  • Existing RDS: To use the existing RDS of Kyvos Manager.

• Select the size of the instance type from the Repository Instance Type.

External Repository for Kyvos 

• Select this option to create a PostgreSQL-based repository using the Amazon RDS. Otherwise, the pre-bundled PostgreSQL-based Kyvos repository will be used.

• Select the size of the instance type from the Repository Instance Type.

Output S3 Storage 

Select this option to create an S3 storage bucket for Kyvos output.

Bucket Name Prefix

Compute Cluster Type

Select one of the following options to select a compute cluster type engine to process semantic model.

• Kyvos Native

• EMR

• Databricks

Kyvos Native Type

This list is displayed only when you select the Kyvos Native option.

Select one of the compute engines to process semantic model.

• Kubernetes

• Query Engine

Kyvos cluster and Databricks cluster will be in same network or in peered network

Select this checkbox if your Kyvos cluster nodes and Databricks cluster are either in the same network or shared using VPC peering. 

NOTE: If this checkbox is not selected, an API gateway method will be created and used for communication between the Databricks cluster and Kyvos BI Server. 

Use Custom Image

(please see the note below) 

Select this option to use a custom image of your environment for deploying the Kyvos cluster. Custom images can be used to bootstrap deployment tasks like preloading applications, application configurations, and other configurations.   

You will have to provide the following information at the time of deploying the cluster through the CloudFormation.

• AmiDefaultUserName: Enter the AMI default logged-in Linux user

• AmiId: Enter the AMI ID to be used for launching Kyvos Instances .  

NOTE: To use your custom image, templates named oe_custom.json , qe_custom.json, km_custom.json, and base_custom.json must be available at the $KYVOS_MANAGER_HOME/cloud/ aws /templates/ location.  

If you are not using a custom image, the cluster will be deployed with Amazon Linux 2.

Deploy with enhanced security

Enhanced security ensures passwords are stored in the AWS Secrets Manager and not in the Kyvos configuration files.

Create resources for managing credit and usage information

Creates or supports the credit-based license support in Kyvos.