Feature Overview (aka. Goal Summary)  

The CEO hardcodes defragmentation exceptions for non-HA topologies (e.g., SNO). The assumption does not fit all customer using non-HA topologies. To work on a fix that adapt to customers needs while maintaining long term reliability of the platform:

• Consider exposing a toggleable option on the CR for nonHA use cases
• Explore the option for conditional defragmentation (e.g., automatic defragmentation when db fragmentation is above certain threshold and db size is over certain size – reference OCPSTRAT-319)

[1] https://github.com/openshift/cluster-etcd-operator/blob/master/pkg/operator/defragcontroller/defragcontroller.go#L106-L110

[2] https://issues.redhat.com/browse/OCPSTRAT-319 

Goals (aka. expected user outcomes)

Provide a supported option to enable/disable defragmentation process on nonHA use cases.

This should align with work on:
https://issues.redhat.com/browse/OCPSTRAT-319 
 

Requirements (aka. Acceptance Criteria):

Because of impact etcd defragmentation can cause on non-HA deployments like SNO, the cluster-admin should retain the ability to completely disable defragmentation.