Create a Cluster

Create a physical cluster to add independent compute and cache capacity to a SaaS or BYOC warehouse. Clusters in the same warehouse share stored data but isolate compute and cache resources.

1. Before you begin

• You need permission to manage compute resources in the warehouse.
• You can create a cluster when the warehouse is Running or Stopped. Creating a cluster automatically starts a stopped warehouse.
• Creating a cluster can incur compute and cache charges. Ensure that your organization has an active deduction channel or sufficient balance. See Billing.

For a BYOC warehouse, your cloud provider bills the infrastructure resources in your cloud account. VeloDB Cloud separately charges the compute service fee for the cluster.

2. Configure the cluster

1. Log in to the VeloDB Cloud console.

2. In the upper-left corner, select the warehouse where you want to create the cluster.

3. In the left navigation pane, under Compute, click Clusters.

4. Click Physical Clusters.

5. Click Add New Cluster. The Create cluster page opens.

6. Under Cluster Setup, configure the following settings:

   Setting	Description
   Cluster Name	Enter a unique name that starts with a letter and contains up to 32 characters. Use only letters, digits, or underscores. Do not use consecutive underscores. Names are compared case-insensitively across physical and virtual clusters in the warehouse.
   Billing Method	Keep On-Demand with Hourly billing selected. VeloDB Cloud measures resource usage by the minute and generates a bill every hour.

7. Under Resources, configure the following settings:

   Setting	Description
   Availability Zone	If this setting appears, select an availability zone. If you plan to create a virtual cluster, select a different availability zone for each physical cluster.
   Memory per vCPU	Select R4 for 4 GB RAM per vCPU for general-purpose workloads, or R8 for 8 GB RAM per vCPU for memory-intensive workloads.
   Cache Storage Type	Select EBS to set the cache capacity now and expand it later. For BYOC warehouses on AWS only, you can also select Local NVMe SSD for high-performance cache with fixed capacity.
   Node Size	Select the vCPU and RAM capacity for each node. The RAM capacity depends on the selected Memory per vCPU option.
   Cache per Node	For EBS, set the cache capacity for each node. Local NVMe SSD has fixed cache capacity.
   Number of Nodes	Set the number of nodes in the cluster.
   Auto-suspend	For an eligible cluster, enable this setting to suspend the cluster after 1-1440 minutes of inactivity. A new workload resumes the cluster automatically.

   Total cluster vCPU, RAM, and cache capacity depend on the per-node configuration and the number of nodes. Use the options and ranges available in the console for your warehouse.

   Local NVMe cache is cleared when the cluster resumes from Auto-suspend or a node is replaced, but is preserved during routine upgrades.

   Note:

   During an active trial, Auto-suspend is enabled for all clusters and fixed at 30 minutes. You cannot edit or disable it. Resume a suspended cluster manually before running workloads.

8. Under Summary, verify the cluster name, billing method, availability zone (if shown), resource configuration, and auto-suspend settings. Confirm that the total compute and cache capacity match your requirements.

9. Review Estimated Cost. The estimate excludes warehouse-level storage charges. For a BYOC warehouse, it also excludes cloud-provider infrastructure charges.

10. Click Create.

3. Verify the cluster

After you click Create, the console opens the cluster list. The new cluster has the Creating status and is ready when its status changes to Running. The cluster is listed under Physical Clusters.

If the status changes to CreateFailed, review the error shown in the console. Correct any account, capacity, or configuration issue before creating the cluster again. If the console does not provide an actionable resolution, contact VeloDB Cloud Support.

Troubleshoot cluster creation

Issue	Resolution
Add New Cluster is unavailable	Confirm that you can manage compute resources and that the warehouse status is Running or Stopped.
Create is unavailable	Complete all required fields and select an available compute and cache configuration.
The console reports insufficient balance	Add funds or configure another supported deduction channel before retrying.

Next steps

• View and manage clusters.
• Scale a cluster as workload demand changes.
• Manage clusters with Terraform.