Workload cluster not resumed during upgrades from sylva 1.3 to 1.4

Summary

During an upgrade from sylva 1.3 to 1.4, workload cluster was not upaused while it was being upgraded:

  ~/my-deployment$ sylvactl status                                                                                                                        
Following resources are not ready:                                                                                                                                                     
IDENTIFIER                                                                        STATUS     REASON              MESSAGE                                                               
Kustomization/test-workload-cluster/cluster                                       InProgress                     Running health checks for revision 1.3.10@sha256:6a4cc1faebfc6e0708515
bb53020327a84caad449929fe142fd64999a521682f with a timeout of 30s                                                                                                                      
╰┄╴HelmRelease/test-workload-cluster/cluster                                      Ready                          Resource is Ready                                                     
   ├┄╴Cluster/test-workload-cluster/wl-cluster                                    InProgress                     Cluster generation is 3, but latest observed generation is 2          
   ┆  ├┄╴MachineDeployment/test-workload-cluster/wl-cluster-md0                   InProgress                     MachineDeployment generation is 2, but latest observed generation is 1
   ┆  ┆  ╰┄╴MachineSet/test-workload-cluster/wl-cluster-md0-pphsn                 InProgress                     MachineSet generation is 2, but latest observed generation is 1       
   ┆  ┆     ├┄╴Machine/test-workload-cluster/wl-cluster-md0-pphsn-z8p7r           Ready                          Resource is Ready                                                     
   ┆  ┆     ┆  ╰┄╴Metal3Machine/test-workload-cluster/wl-cluster-md0-pphsn-z8p7r  InProgress                                                                                           
   ┆  ┆     ┆     ╰┄╴┬┄┄[Conditions]                                                                                                                                                   
   ┆  ┆     ┆        ├┄╴Ready                                                     False      Metal3MachinePaused                                                                       
   ┆  ┆     ┆        ├┄╴AssociateBMH                                              False      Metal3MachinePaused                                                                       
   ┆  ┆     ┆        ├┄╴KubernetesNodeReady                                       True                                                                                                 
   ┆  ┆     ┆        ╰┄╴Metal3DataReady                                           True                                                                                                 
   ┆  ┆     ╰┄╴Machine/test-workload-cluster/wl-cluster-md0-pphsn-n26ss           Ready                          
[...]

This was caused by this condition that was not satisfied as sylva-unit-status had not reconciled in sylva 1.3.

This is a corner case, as sylva-units is expected to be in a stable state prior to launch the upgrade.

But as this condition is only an optimisation, we could run it systematically to handle such cases.