feat(channel-monitor): redesign user dashboard as card grid

Reference check-cx UI: INTELLIGENCE MONITOR hero + 3-column card grid
with 60-point timeline bars.

Backend:
- Add PrimaryPingLatencyMs + Timeline[60] to UserMonitorView
- ListRecentHistoryForMonitors: batch CTE + ROW_NUMBER() window query
- indexLatestByModel / indexAvailabilityByModel helpers

Frontend:
- 7 new components: ProviderIcon, MonitorMetricPair, MonitorAvailabilityRow,
  MonitorTimeline, MonitorHero, MonitorCard, MonitorCardGrid
- ChannelStatusView 381→~180 lines (delegated to subcomponents)
- AbortController reload concurrency protection
- HSL 0-120° availability color mapping
- Replace emoji with Icon component (bolt / globe)
- i18n: monitorCommon.* shared namespace, channelStatus.hero.*

Bump VERSION to 0.1.114.24

backend/internal/service/channel_monitor_service.go

@@ -38,6 +38,9 @@ type ChannelMonitorRepository interface {
 	// 批量聚合（admin/user list 用，避免 N+1）
 	ListLatestForMonitorIDs(ctx context.Context, ids []int64) (map[int64][]*ChannelMonitorLatest, error)
 	ComputeAvailabilityForMonitors(ctx context.Context, ids []int64, windowDays int) (map[int64][]*ChannelMonitorAvailability, error)
+	// ListRecentHistoryForMonitors 批量取多个 monitor 各自主模型（primaryModels[monitorID]）最近 perMonitorLimit 条历史。
+	// 返回的 entry 已按 checked_at DESC 排序（最新在前），不含 message 字段。
+	ListRecentHistoryForMonitors(ctx context.Context, ids []int64, primaryModels map[int64]string, perMonitorLimit int) (map[int64][]*ChannelMonitorHistoryEntry, error)
 }
 
 // ChannelMonitorService 渠道监控管理服务。