feat: scheduled backups, availability tracking, PoE power metrics (v0.11.2 Beta)

P1 — Scheduled Backups
- BackupService.createBackup() now accepts a `type` param (manual/scheduled)
- PollerService: cronMatchesNow() parser + runScheduledBackups() + Redis 1h gate
- Settings page: backup schedule toggle + cron expression input with examples

P2 — Uptime/Availability Tracking
- DB: device_availability table (went_offline_at, came_back_online_at, duration_seconds)
- PollerService: inserts outage rows on offline detection, closes them on recovery
- GET /api/metrics/device/:id/availability?range=30d endpoint
- /api/metrics/summary now includes fleet 30-day uptime %
- DeviceDetailPage: 30-day availability card (uptime %, outage count, longest outage)
- DashboardPage HealthBar: Uptime 30d mini-stat

P3 — PoE Power Draw Metrics
- DeviceCollector.collectPoePower() queries /interface/ethernet/poe/monitor for switches
- Writes poe_power measurement to InfluxDB (tags: device_id, port; fields: watts, current_ma, voltage_v)
- GET /api/metrics/device/:id/poe endpoint (current per-port + 1h history)
- SwitchPortDiagram: amber watt badge on PoE-active ports
- HardwareTab: PoE budget section showing active ports and total watts

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

frontend/src/types/index.ts

@@ -230,6 +230,21 @@ export interface MetricsSummary {
   devices: { total: number; online: number; offline: number };
   clients: { total: number; active: number };
   alerts: { critical: number; warning: number };
+  availability?: { fleetUptimePct30d: number };
+}
+
+export interface DeviceAvailability {
+  uptimePct: number;
+  totalOutages: number;
+  longestOutageSec: number;
+  totalOutageSec: number;
+  outages: {
+    id: number;
+    went_offline_at: string;
+    came_back_online_at: string | null;
+    duration_seconds: number | null;
+  }[];
+  range: string;
 }
 
 export interface TimeSeriesPoint {