Fix Tenant PodInformer for Tenant Status and remove recurrent job

[cesnietor]

Use PodInformer to Update Tenant status.
Previous code had both an informer and a recurrent job (every 3 min) but it wasn't working great so this changes so that it updates the tenant status every time there's a change on the pods.
Before informer was reacting on all pods and then filtering them when handling the event, now it only reacts on pods that match the label selector.

Changes:

• Use custom informer so that it can react only on pods that contain the Tenant labels
• Requeues event if tenant is not in green status
• Remove recurrent health job
• Updates al the places where we had v1.min.io/tenant with miniov2.TenantLabel

Test Steps

• Add a Tenant
• Do some changes to the Pod
• Delete the pod
• Tenant Status should be updated on both changes (use node cordon if needed)

[ramondeklein]

Great improvement too to remove the "v1.min.io/tenant" literals. Only a single remark about a function that implies to do more than it can and inlining it may be a better alternative.

[pjuarezd]

this might need some more work, shared findings with @cesnietor wiating for his opinion before merge

[pjuarezd]

this might need some more work, shared findings with @cesnietor wiating for his opinion before merge

All good now, tested and doubled tested

[cesnietor]

this might need some more work, shared findings with @cesnietor wiating for his opinion before merge

for documenting purposes:
Pedro noticed that pod changes was not enough to trigger an update and get the latest status cause the service might take some time to be ready (green) and that wouldn't trigger another health check so we now enqueue again if the service is not ready. The queue already has a backoff so that takes care of the load.