fix(controllers): do not reconcile status updates

[jlevesy]

What does this PR do?

This PR places an event filter when setting up that filters out updates to the status of all managed resources by this controller. We only want to reconcile the dashboards only if there's a change in the spec of the object, not the status.

I'm not 100% sure of myself but I suspect this creates an infinite loop of events: as it is, when the first reconciliation is done, dark sets the status to OK, which triggers another update and another reconciliation, and another status update... and so on and so forth 😅

Happy to chat about this.

Reference: kubernetes-sigs/kubebuilder#1103 and kubernetes-sigs/kubebuilder#618

[jlevesy]

CI failure looks unrelated. I'll take a stab at fixing it.

[codecov-commenter]

Codecov Report

Merging #233 (653c817) into master (675523d) will decrease coverage by 0.07%.
Report is 4 commits behind head on master.
The diff coverage is 16.66%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #233      +/-   ##
==========================================
- Coverage   40.99%   40.93%   -0.07%     
==========================================
  Files          39       39              
  Lines        2583     2587       +4     
==========================================
  Hits         1059     1059              
- Misses       1484     1488       +4     
  Partials       40       40              

Files	Coverage Δ
...nternal/pkg/controllers/alertmanager_controller.go	0.00% <0.00%> (ø)
internal/pkg/controllers/apikey_controller.go	0.00% <0.00%> (ø)
internal/pkg/controllers/datasource_controller.go	0.00% <0.00%> (ø)
...nal/pkg/controllers/grafanadashboard_controller.go	0.00% <0.00%> (ø)
internal/pkg/converter/gauge.go	55.00% <50.00%> (ø)

[K-Phoen]

Thanks!