feat: expose operator metrics

[jaideepr97]

What type of PR is this?

Uncomment only one /kind line, and delete the rest.
For example, > /kind bug would simply become: /kind bug

/kind bug
/kind chore
/kind cleanup
/kind failing-test
/kind enhancement
/kind documentation
/kind code-refactoring

What does this PR do / why we need it:
This PR exposes custom metrics that track some operator performance vectors to port 8080 of the operator deployment. It adds logic to track when new argo-cd instances are created, deleted, and transition between phases and exposes the counts as metrics. It also exposes a new metric to track reconciliation times for each instance being managed. This PR also adds a service and and servicemonitor to expose the metrics to prometheus.

Have you updated the necessary documentation?

• Documentation update is required by this PR.
• Documentation has been updated.

Which issue(s) this PR fixes:

Fixes #?

How to test changes / Special notes to the reviewer:

1. Run make install deploy
2. Verify that service for metrics is created
3. create a basic argo-cd instance
4. Run kubectl port-forward <argocd-operator-pod-name> 8080:8080 -n argocd-operator-system
5. Go to localhost:8080 and verify that you can see the following metrics included:

# HELP active_argocd_instances_by_phase Number of active argocd instances by phase
# TYPE active_argocd_instances_by_phase gauge
active_argocd_instances_by_phase{phase="Available"} 1
active_argocd_instances_by_phase{phase="Pending"} 0
# HELP active_argocd_instances_total Total number of active argocd instances
# TYPE active_argocd_instances_total gauge
active_argocd_instances_total 1

...

# HELP controller_runtime_reconcile_time_seconds_per_instance Length of time per reconciliation per instance
# TYPE controller_runtime_reconcile_time_seconds_per_instance histogram
controller_runtime_reconcile_time_seconds_per_instance_bucket{namespace="default",le="0.005"} 0
controller_runtime_reconcile_time_seconds_per_instance_bucket{namespace="default",le="0.01"} 1
controller_runtime_reconcile_time_seconds_per_instance_bucket{namespace="default",le="0.025"} 1