RFC: add System replicas for memcachedResources & sphinxSpec in reconciliable parameters

[ghijsma]

Using:

Openshift 4.13 k8s: v1.26.9+aa37255 ArgoCD v2.9.2+c5ea5c4 3scale-operator: 0.10.7 provided by Red Hat 3Scale APIManager: v2.13

Issue: We have a multicluster setup in Openshift 4. For this to work during downtime we should be able to failover our APIManager from one cluster to the other. 3Scale cannot run on both clusters because of issues with the shared databases. Therefore it should be possible for us to change the replica's in all the components.

This is not possible for these two components in the current operator:

• system:momcachedResources
• system:sphinxSpec

System replicas

Solution: Add support for these two components in the operator so that we can change the number of replica's for them in our gitops. Making it possible for us to fully scale down to "0" one APIManager on one cluster and scale up the other on the other cluster.

[briangallagher]

Hi @ghijsma ,

The Sphinx component is removed in 2.14, no plans to backport changes here.

For memcache, you could add the deploymentconfig to git and scale the replica there to 0. Would that work? It was not added to the APIManager CRD as it doesn't support replication.