Skip to main content
Version: 1.1

Configure Job Clusters in Rift

When defining a Batch Feature View or a Stream Feature View using Rift, you can set the batch_compute parameter using the RiftBatchConfig object.

This configures the instance type of the job used to materialize to the online and/or offline store. This must be set to a valid EC2 instance type.

Example​

from tecton import batch_feature_view, RiftBatchConfig


@batch_feature_view(
    batch_compute=RiftBatchConfig(
        instance_type="m5.2xlarge",
    ),
    ...,
)
def my_feature_view():
    return ...

Instance Type Ratios by Deployment Model​

The Tecton Credits consumed per hour for Rift instance types varies depending on your deployment model:

Hybrid SaaS Deployment​

When Rift runs in the Tecton Control Plane (Hybrid SaaS deployment), it uses Rift-optimized instance type ratios. For detailed ratio information, see Rift Instance Types.

Standard Virtual Private Tecton Deployment​

When Rift runs in the Customer Data Plane (standard Virtual Private Tecton deployment), it uses the same instance type ratios as Spark. For detailed ratio information, see Spark Instance Types.

To understand which deployment model applies to your organization, see Virtual Private Tecton Architecture.

Was this page helpful?