vpa_enable

Description

The vpa_enable parameter installs the Vertical Pod Autoscaler (VPA) on the rack. VPA automatically adjusts CPU and memory requests for services based on observed usage patterns, right-sizing each replica's resource allocation over time.

Default Value

The default value for vpa_enable is false.

Use Cases

Resource Optimization: Automatically right-size service resource requests based on actual usage instead of manual estimation.
Cost Reduction: Avoid over-provisioning CPU and memory by letting VPA recommend or apply optimal values.
Performance Improvement: Prevent under-provisioning that can lead to CPU throttling or OOM kills.

Setting Parameters

To enable VPA on your rack, use the following command:

$ convox rack params set vpa_enable=true -r rackName
Setting parameters... OK

Additional Information

Enabling VPA installs the VPA controller, recommender, and admission controller in the cluster. Once enabled, services can use the scale.vpa section in their convox.yml to configure vertical autoscaling.

See VPA for service configuration details.

Use Cases

Resource Optimization: Automatically right-size service resource requests based on actual usage instead of manual estimation.

Cost Reduction: Avoid over-provisioning CPU and memory by letting VPA recommend or apply optimal values.

Performance Improvement: Prevent under-provisioning that can lead to CPU throttling or OOM kills.