fix: ResourceDeletedTotal metric registered but never incremented

[digiserg]

Summary

ResourceDeletedTotal counter metric is declared and registered in internal/metrics/metrics.go:48 but never incremented anywhere in the codebase. This is dead code that provides misleading observability — the metric exists in Prometheus but always reads zero.

Detailed Requirements

Either:

1. Wire the metric into delete paths across all controllers (increment on successful resource deletion), OR
2. Remove the metric entirely and deregister from the Prometheus collector

Option 1 is preferred if the metric provides operational value.

Location

• internal/metrics/metrics.go:48

Acceptance Criteria

1. Metric is either wired into all controller delete paths OR removed from registration
2. If kept: incremented on successful resource deletion in each controller (AxonOpsServer, all alert controllers)
3. If removed: deregistered from Prometheus collector, no dangling references

Testing Requirements

• If kept: unit test verifies metric increments on successful deletion
• If removed: verify no references remain in codebase
• Grep for any other registered-but-unused metrics

Documentation Requirements

• If kept: document the metric in operator metrics reference
• If removed: no action needed

Dependencies

• None