feat: 2.0.0-alpha.11 — DB-backed service registry + LAN auto-discovery

- monitored_services table (mig 015) replaces config/services.json (now a boot seed)
- owner CRUD over /api/health/services; GET is DB-backed; cron+worker read the DB
- discover.lan worker: pure-Node TCP sweep + HTTP-title probe -> disabled 'discovered'
  candidates (never clobbers curated entries); POST /api/health/discover + GET .../discovered
- dashboard: Scan button + Discovered(N) section with one-click promote

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

lib/cron/index.js

@@ -2,9 +2,9 @@ import cron from 'node-cron';
 import { runSync } from './sync_source_docs.js';
 import { log } from '../log.js';
 import { enqueue } from '../jobs/queue.js';
-import { load } from '../health/registry.js';
 import { checkAll } from '../health/checker.js';
 import * as statusRepo from '../db/repos/service_status.js';
+import * as services from '../db/repos/monitored_services.js';
 
 export function startCron() {
   // Daily at 03:00 local time
@@ -29,7 +29,7 @@ export function startCron() {
   // Keep the two in sync — both rely on lib/health/checker.js as the source of truth.
   cron.schedule('*/1 * * * *', async () => {
     try {
-      const results = await checkAll(load());
+      const results = await checkAll(await services.listEnabled());
       for (const r of results) await statusRepo.upsert(r);
       log.info({ n: results.length }, 'health check complete');
     } catch (e) { log.error({ err: e }, 'health check failed'); }