Deploying a new cluster
You can create and deploy a new cluster. After deploying the cluster, the deployment details are displayed.
You must maintain a minimum balance relative to the clusters being deployed. The minimum balance is calculated as N × $100, where N is the total number of clusters being deployed or already deployed.
To deploy a cluster, perform the following steps.
On the navigation bar, click Clusters.
Click the New button. The New Cluster dialog is displayed.
Select the required Subscription from the list. The current balance of the selected subscription is displayed.
Note
If the current balance for the selected subscription is low, first you must add funds to that subscription to create a new cluster.
Enter the Cluster Name.
Enter Description of the cluster if required.
Select AWS or Azure from the Cloud Provider list where the cluster needs to be created.
Select the geographic location from the Region list where your clusters will be deployed.
Select the required deployment type from the Deployment Type list.
Single-node deployment: Select this option to run all Kyvos services on a single node. It deploys on existing nodes with bundled Postgres. Disaster recovery and high availability are not supported with single-node deployment.
Multi-node deployment: Select this option to run all Kyvos services on separate nodes.
For Azure, select one of the following options for Workload Type cluster:
Development Workload Clusters: This is designed for cost-optimization configurations for Postgres and other services.
Production Workload Clusters: This is designed for production grade configurations for Postgres and other services.
Feature | Development | Production |
Application Gateway | Basic | WAF_V2 |
Flexible Server | Burstable | General purpose |
Select the Instances Size from the list to specify the capacity of a node that impacts its workload capacity.
Note:This sizing option is applicable only to BI Server and Query Engine instances.
The Web Portal, Kyvos Manager, and Dedicated Compute Cluster instance types remain unchanged.
It is applicable to Dedicated Compute, Shared Compute, and Single Node Installer deployments.
For AWS
Small: m6i.2xlarge
Medium: m6i.4xlarge
Large: m6i.8xlarge
For Azure
Small: Standard_D8s_v5
Medium: Standard_D16s_v5
Large: Standard_D32s_v5
Select one of the following options in the Data Processing. This is applicable only when you create a new cluster.
Process on Existing Nodes: Select this option to process data on the same nodes as Query Engines and other services. Saves cost, but performance may be limited for heavy workloads.
Process on Dedicated Nodes: Select this option to add separate nodes on demand for data processing. Ensures higher performance for large or intensive workloads.
NOTE: Kyvos does not support Kubernetes (no-Spark) and EMR (Spark-based) deployments.
Select the number of Query Engine(s) Instances allocated to the cluster.
Enter the IP range in CIDR format for the cluster's services and nodes in the CIDR for Services box.
Feature | Development | Production |
Application Gateway | Basic | WAF_V2 |
Flexible Server | Burstable | General purpose |
Note
CIDR range starts from 10.0.0.0/16 to 10.255.255.255/16. You will receive an auto-assigned CIDR based on availability, but it can be modified within the permitted range if needed.
Select the System Availability checkbox.
If this is enabled, two Web Portals, two BI Server nodes, and a minimum of two Query Engines are deployed across multiple availability zones. The Query Engines instances will be auto selected as 2. You can also modify the number of Query Engine instances as needed. The maximum Query Engine can be selected as 5.
If this is disabled, the deployment consists of one Web Portal, one BI Server node, and the specified number of Query Engines—all in a single availability zone to optimize cost. You can specify the number of Query Engines (QEs) between 1 to 5 to be provisioned during deployment based on workload and performance requirements.
NOTE: With high availability enabled, the system remains operational during availability zone failures.
Select the Disaster Recovery checkbox to enable or disable backup and disaster recovery services for the region.
Select the Backup Region from the list. This is a geographical location where data, metadata, and configuration will be stored as a backup to facilitate cluster restoration on case of disaster.
Enter the CIDR for DR. Read the onscreen information.
In the Login Credentials for Kyvos Manager and Kyvos Web Portal section, all details are prepopulated. You just need to enter a Password you want to set.
Retype the password to confirm. Ensure that the user name (Admin) and password will remain the same for the Kyvos Manager and Kyvos Web Portal.
Click Review and Deploy to create and deploy cluster. The Review and Deploy dialog is displayed.
Click Confirm. Once the cluster is deployed, the Kyvos SaaS Portal will give you access to Kyvos Manager and Kyvos Web Portal. User must provide the password setup for cluster to log into these portals.
Note
The Backup region must be AWS Region.
The password must be:
Minimum of 8 characters
At least one capital letter
At least one special character
At least one number
For example, PassW0rd@100
Once the cluster is deployed, you can view the deploying details.