Automatic scaling resources for Assistant Builder in watsonx Orchestrate

The watsonx™ Orchestrate supports auto scaling to automatically adjust resources based on workload demand. You can enable or disable auto scaling by using the oc patch command.

Note: Auto scaling is supported only for Assistant Builder services. Other watsonx Orchestrate services do not support auto scaling.

Enable auto scaling

To enable auto scaling for your watsonx Orchestrate instance, run:

oc patch wo wo --namespace=${PROJECT_CPD_INST_OPERANDS} --type='merge' -p '{"spec":{"autoScaleConfig":true}}'

The oc patch command sets the autoScaleConfig field to true, allowing watsonx Orchestrate to scale automatically according to system utilization.

Disable auto scaling

To disable auto scaling and return to fixed or manual scaling, run:

oc patch wo wo --namespace=${PROJECT_CPD_INST_OPERANDS} --type='merge' -p '{"spec":{"autoScaleConfig":false}}'

Setting the autoScaleConfig value to false prevents watsonx Orchestrate from automatically adjusting resource usage.

Verify the setting

To verify the current auto scaling configuration, run:

oc get wo wo --namespace=${PROJECT_CPD_INST_OPERANDS} -o yaml | grep autoScaleConfig

This displays the current auto scaling value (true or false) in the watsonx Orchestrate custom resource.