Add usage metrics collector for Splunk ingestion

[boubz1623]

Summary

• Implements a data collector that aggregates per-workspace and per-tenant usage metrics (components, relationships, recommendations, requirements) from internal CATIO services
• Emits structured JSON events to ~/.config/kportforward/usage-metrics.log for Splunk ingestion via the existing stdout logging pipeline
• Background collection integrated into kportforward — triggers automatically once services are healthy, runs once per day with idempotency state tracking

What's included

• New kportforward collect --once command for manual or scripted runs
• gRPC clients wired to architecture-inventory, recommendations, and requirements services using local proto definitions from catio-tech/protos
• HTTP client for workspace discovery via environment service
• JSON-only output file — diagnostic text is discarded so Splunk receives clean JSON lines
• Remote config fallback so the collector works even when the remote config predates this feature

CI status

CI is currently failing because this public repo depends on catio-tech/protos (private) via a replace directive in go.mod. The GitHub Actions runner does not have access to it. The code works correctly locally. This will be resolved when the collector is moved to a private repo as discussed.

Test plan

• Run .\bin\kportforward.exe collect --once --force 2>$null and verify JSON events are emitted to stdout
• Run .\bin\kportforward.exe and verify ~/.config/kportforward/usage-metrics.log is created and contains only JSON lines after services come up
• Verify component_count and relationship_count are non-zero
• Run go test ./... — all tests pass