CLPR-13: Observability

## Epic CLPR-13: Observability

**Parent:** hiero-ledger/roadmap#48
**Repo:** hiero-ledger/hiero-consensus-node + Besu plugin repo

### Objective

Monitoring, alerting, and diagnostic tooling for CLPR on both platforms.

### Expected Functionality

1. **Hiero metrics** — Prometheus: sync ops (count, latency, success/failure), bundle processing (messages/bundle, verification time), queue depth, proof construction time, endpoint reputation, Connector economics.
2. **Besu metrics** — Same categories plus: gas usage, tx confirmation latency, sync committee freshness.
3. **Alerting rules** — Connection HALTED, queue near max, Connector balance low, sync failure rate high, proof staleness, verifier code hash change.
4. **Dashboards** — Grafana templates: overview (all Connections), per-Connection detail, economic view.
5. **Diagnostics** — Message trace (source → destination → response), queue inspection, Connection health check.

### Dependencies

- CLPR-4 (Hiero Endpoint metrics)
- CLPR-8 (Besu Endpoint metrics)

### Risk

**Low** — Additive, doesn't affect correctness. Value increases with operational experience.

Start by reading the dependency issues CLPR-4 and CLPR-8, then map the Hiero and Besu plugin entry points described in this epic. The work is done when the listed metrics, alerts, Grafana dashboards, and diagnostic tools are defined and implemented for both platforms without affecting correctness.

Written by the indexing model from the issue text.