refactor(ops): 移除duration相关告警指标，简化监控配置

主要改动：
- 移除 p95_latency_ms 和 p99_latency_ms 告警指标类型
- 移除配置中的 latency_p99_ms_max 阈值设置
- 简化健康分数计算（移除latency权重，重新归一化SLA和错误率）
- 移除duration相关的诊断规则和阈值检查
- 统一术语：延迟 → 请求时长
- 保留duration数据展示，但不再用于告警判断
- 聚焦TTFT作为主要的响应速度告警指标

影响范围：
- Backend: handler, service, models, tests
- Frontend: API types, i18n, components

frontend/src/i18n/locales/en.ts

@@ -1887,7 +1887,7 @@ export default {
       totalRequests: 'Total Requests',
       avgQps: 'Avg QPS',
       avgTps: 'Avg TPS',
-      avgLatency: 'Avg Latency',
+      avgLatency: 'Avg Request Duration',
       avgTtft: 'Avg TTFT',
       exceptions: 'Exceptions',
       requestErrors: 'Request Errors',
@@ -1899,7 +1899,7 @@ export default {
       errors: 'Errors',
       errorRate: 'error_rate:',
       upstreamRate: 'upstream_rate:',
-      latencyDuration: 'Latency (duration_ms)',
+      latencyDuration: 'Request Duration (ms)',
       ttftLabel: 'TTFT (first_token_ms)',
       p50: 'p50:',
       p90: 'p90:',
@@ -1919,7 +1919,7 @@ export default {
       failedToLoadData: 'Failed to load ops data.',
       failedToLoadOverview: 'Failed to load overview',
       failedToLoadThroughputTrend: 'Failed to load throughput trend',
-      failedToLoadLatencyHistogram: 'Failed to load latency histogram',
+      failedToLoadLatencyHistogram: 'Failed to load request duration histogram',
       failedToLoadErrorTrend: 'Failed to load error trend',
       failedToLoadErrorDistribution: 'Failed to load error distribution',
       failedToLoadErrorDetail: 'Failed to load error detail',
@@ -1927,7 +1927,7 @@ export default {
       tpsK: 'TPS (K)',
       top: 'Top:',
       throughputTrend: 'Throughput Trend',
-      latencyHistogram: 'Latency Histogram',
+      latencyHistogram: 'Request Duration Histogram',
       errorTrend: 'Error Trend',
       errorDistribution: 'Error Distribution',
       // Health Score & Diagnosis
@@ -1973,14 +1973,7 @@ export default {
         memoryHigh: 'Memory usage elevated ({usage}%)',
         memoryHighImpact: 'Memory pressure is high, needs attention',
         memoryHighAction: 'Monitor memory trends, check for memory leaks',
-        // Latency diagnostics
-        latencyCritical: 'Response latency critically high ({latency}ms)',
-        latencyCriticalImpact: 'User experience extremely poor, many requests timing out',
-        latencyCriticalAction: 'Check slow queries, database indexes, network latency, and upstream services',
-        latencyHigh: 'Response latency elevated ({latency}ms)',
-        latencyHighImpact: 'User experience degraded, needs optimization',
-        latencyHighAction: 'Analyze slow request logs, optimize database queries and business logic',
-        ttftHigh: 'Time to first byte elevated ({ttft}ms)',
+        ttftHigh: 'Time to first token elevated ({ttft}ms)',
         ttftHighImpact: 'User perceived latency increased',
         ttftHighAction: 'Optimize request processing flow, reduce pre-processing time',
         // Error rate diagnostics
@@ -2020,7 +2013,7 @@ export default {
         context: 'Context',
         status: 'Status',
         message: 'Message',
-        latency: 'Latency',
+        latency: 'Request Duration',
         action: 'Action',
         noErrors: 'No errors in this window.',
         grp: 'GRP:',
@@ -2049,7 +2042,7 @@ export default {
         basicInfo: 'Basic Info',
         platform: 'Platform',
         model: 'Model',
-        latency: 'Latency',
+        latency: 'Request Duration',
         ttft: 'TTFT',
         businessLimited: 'Business Limited',
         requestPath: 'Request Path',
@@ -2398,8 +2391,6 @@ export default {
         metricThresholdsHint: 'Configure alert thresholds for metrics, values exceeding thresholds will be displayed in red',
         slaMinPercent: 'SLA Minimum Percentage',
         slaMinPercentHint: 'SLA below this value will be displayed in red (default: 99.5%)',
-        latencyP99MaxMs: 'Latency P99 Maximum (ms)',
-        latencyP99MaxMsHint: 'Latency P99 above this value will be displayed in red (default: 2000ms)',
         ttftP99MaxMs: 'TTFT P99 Maximum (ms)',
         ttftP99MaxMsHint: 'TTFT P99 above this value will be displayed in red (default: 500ms)',
         requestErrorRateMaxPercent: 'Request Error Rate Maximum (%)',
@@ -2458,7 +2449,7 @@ export default {
       tooltips: {
         totalRequests: 'Total number of requests (including both successful and failed requests) in the selected time window.',
         throughputTrend: 'Requests/QPS + Tokens/TPS in the selected window.',
-        latencyHistogram: 'Latency distribution (duration_ms) for successful requests.',
+        latencyHistogram: 'Request duration distribution (ms) for successful requests.',
         errorTrend: 'Error counts over time (SLA scope excludes business limits; upstream excludes 429/529).',
         errorDistribution: 'Error distribution by status code.',
         goroutines:
@@ -2473,7 +2464,7 @@ export default {
         sla: 'Service Level Agreement success rate, excluding business limits (e.g., insufficient balance, quota exceeded).',
         errors: 'Error statistics, including total errors, error rate, and upstream error rate.',
         upstreamErrors: 'Upstream error statistics, excluding rate limit errors (429/529).',
-        latency: 'Request latency statistics, including p50, p90, p95, p99 percentiles.',
+        latency: 'Request duration statistics, including p50, p90, p95, p99 percentiles.',
         ttft: 'Time To First Token, measuring the speed of first byte return in streaming responses.',
         health: 'System health score (0-100), considering SLA, error rate, and resource usage.'
       },