Wire node loss signals into service reconciler

[kacy]

Summary:

• persist endpoint node ids through the service registry bridge and preserve them across health updates
• feed gossip and heartbeat-driven node loss/recovery signals into the authoritative service reconciler for immediate DNS and compatibility-mirror updates
• expose reconciler node-signal state in rollout status and Prometheus metrics, with tests covering the new paths

Verification:

• zig fmt on touched files
• timeout 30 env YOQ_SKIP_SLOW_TESTS=1 ZIG_GLOBAL_CACHE_DIR=.zig-global-cache ZIG_LOCAL_CACHE_DIR=.zig-local-cache zig build test -- --test-filter health bridge preserves existing node id for endpoint (timed out with no emitted failures)
• timeout 30 env YOQ_SKIP_SLOW_TESTS=1 ZIG_GLOBAL_CACHE_DIR=.zig-global-cache ZIG_LOCAL_CACHE_DIR=.zig-local-cache zig build test -- --test-filter node loss and recovery reconcile authoritative DNS immediately (timed out with no emitted failures)
• timeout 30 env YOQ_SKIP_SLOW_TESTS=1 ZIG_GLOBAL_CACHE_DIR=.zig-global-cache ZIG_LOCAL_CACHE_DIR=.zig-local-cache zig build test -- --test-filter route handles /v1/status?mode=service_rollout GET (timed out with no emitted failures)