feat: Hub monitoring takeover — event system, dead man's switch, notifications (v0.3.0)

Replace external Healthchecks.io with Hub-native monitoring. New events
table + /api/v1/event endpoint for structured events from controllers.
Staleness checker (60s) detects unresponsive nodes. Backup deadline
checker (daily 05:00) catches missed backups. Notification dispatcher
sends operator (English) + customer (Hungarian) emails via Resend with
per-event cooldowns. Event timeline on customer page, dashboard badges.
Config form deprecates Monitoring UUIDs section.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

hub/internal/web/templates/config_form.html

@@ -96,8 +96,9 @@
                 </div>
             </details>
 
-            <details class="card" {{if index .Overrides "monitoring"}}open{{end}}>
-                <summary><h2 style="display:inline">Monitoring UUIDs</h2></summary>
+            <details class="card">
+                <summary><h2 style="display:inline">Monitoring UUIDs</h2> <span class="severity-badge severity-warning" style="font-size: 0.7em; vertical-align: middle;">Legacy</span></summary>
+                <p class="text-muted" style="margin: 0.5rem 0;">Healthchecks ping UUIDs are deprecated. Monitoring is now handled natively by the Hub event system. These fields are kept for backward compatibility with older controllers.</p>
                 <div class="form-grid" style="margin-top: 1rem;">
                     {{$uuids := ""}}
                     {{with .Overrides}}{{with index . "monitoring"}}{{with index . "ping_uuids"}}{{$uuids = .}}{{end}}{{end}}{{end}}