feat(nip66): add RelayMonitorWorker cluster worker and probe scheduler #695

Summary

Add a new cluster worker (WORKER_TYPE=relay-monitor) that runs NIP-66 probes on a configurable interval and persists the latest probe snapshot for the admin API and event publisher.
Part of #667.

Problem

The shared probe engine (#669) can measure external relay health, but nothing schedules probes, stores results, or wires into the cluster worker model Nostream already uses (maintenance, static-mirroring).