fix(ops): 修复告警状态验证和错误处理逻辑

- 增强告警事件状态验证，添加合法状态值检查
- 移除重试逻辑中的遗留字段赋值
- 修正仓库不可用时的错误类型
- 格式化测试文件代码

backend/internal/handler/admin/ops_alerts_handler.go

@@ -544,8 +544,14 @@ func (h *OpsHandler) ListAlertEvents(c *gin.Context) {
 		}
 	}
 
-	// Cursor pagination
-	if rawTS := strings.TrimSpace(c.Query("before_fired_at")); rawTS != "" {
+	// Cursor pagination: both params must be provided together.
+	rawTS := strings.TrimSpace(c.Query("before_fired_at"))
+	rawID := strings.TrimSpace(c.Query("before_id"))
+	if (rawTS == "") != (rawID == "") {
+		response.BadRequest(c, "before_fired_at and before_id must be provided together")
+		return
+	}
+	if rawTS != "" {
 		ts, err := time.Parse(time.RFC3339Nano, rawTS)
 		if err != nil {
 			if t2, err2 := time.Parse(time.RFC3339, rawTS); err2 == nil {
@@ -557,7 +563,7 @@ func (h *OpsHandler) ListAlertEvents(c *gin.Context) {
 		}
 		filter.BeforeFiredAt = &ts
 	}
-	if rawID := strings.TrimSpace(c.Query("before_id")); rawID != "" {
+	if rawID != "" {
 		id, err := strconv.ParseInt(rawID, 10, 64)
 		if err != nil || id <= 0 {
 			response.BadRequest(c, "Invalid before_id")

backend/internal/repository/ops_repo.go

@@ -925,9 +925,13 @@ func buildOpsErrorLogsWhere(filter *service.OpsErrorLogFilter) (string, []any) {
 	// ops_error_logs stores client-visible error requests (status>=400),
 	// but we also persist "recovered" upstream errors (status<400) for upstream health visibility.
 	// By default, keep list endpoints scoped to unresolved records if the caller didn't specify.
-	if filter != nil && filter.Resolved == nil {
+	resolvedFilter := (*bool)(nil)
+	if filter != nil {
+		resolvedFilter = filter.Resolved
+	}
+	if resolvedFilter == nil {
 		f := false
-		filter.Resolved = &f
+		resolvedFilter = &f
 	}
 	// Keep list endpoints scoped to client errors unless explicitly filtering upstream phase.
 	if phaseFilter != "upstream" {
@@ -967,8 +971,8 @@ func buildOpsErrorLogsWhere(filter *service.OpsErrorLogFilter) (string, []any) {
 		args = append(args, source)
 		clauses = append(clauses, "LOWER(COALESCE(error_source,'')) = $"+itoa(len(args)))
 	}
-	if filter.Resolved != nil {
-		args = append(args, *filter.Resolved)
+	if resolvedFilter != nil {
+		args = append(args, *resolvedFilter)
 		clauses = append(clauses, "COALESCE(resolved,false) = $"+itoa(len(args)))
 	}
 	if len(filter.StatusCodes) > 0 {

backend/internal/service/ops_alert_evaluator_service.go

@@ -206,7 +206,7 @@ func (s *OpsAlertEvaluatorService) evaluateOnce(interval time.Duration) {
 			continue
 		}
 
-		scopePlatform, scopeGroupID := parseOpsAlertRuleScope(rule.Filters)
+		scopePlatform, scopeGroupID, scopeRegion := parseOpsAlertRuleScope(rule.Filters)
 
 		windowMinutes := rule.WindowMinutes
 		if windowMinutes <= 0 {
@@ -239,7 +239,7 @@ func (s *OpsAlertEvaluatorService) evaluateOnce(interval time.Duration) {
 			// Scoped silencing: if a matching silence exists, skip creating a firing event.
 			if s.opsService != nil {
 				platform := strings.TrimSpace(scopePlatform)
-				region := (*string)(nil)
+				region := scopeRegion
 				if platform != "" {
 					if ok, err := s.opsService.IsAlertSilenced(ctx, rule.ID, platform, scopeGroupID, region, now); err == nil && ok {
 						continue
@@ -370,9 +370,9 @@ func requiredSustainedBreaches(sustainedMinutes int, interval time.Duration) int
 	return required
 }
 
-func parseOpsAlertRuleScope(filters map[string]any) (platform string, groupID *int64) {
+func parseOpsAlertRuleScope(filters map[string]any) (platform string, groupID *int64, region *string) {
 	if filters == nil {
-		return "", nil
+		return "", nil, nil
 	}
 	if v, ok := filters["platform"]; ok {
 		if s, ok := v.(string); ok {
@@ -403,7 +403,15 @@ func parseOpsAlertRuleScope(filters map[string]any) (platform string, groupID *i
 			}
 		}
 	}
-	return platform, groupID
+	if v, ok := filters["region"]; ok {
+		if s, ok := v.(string); ok {
+			vv := strings.TrimSpace(s)
+			if vv != "" {
+				region = &vv
+			}
+		}
+	}
+	return platform, groupID, region
 }
 
 func (s *OpsAlertEvaluatorService) computeRuleMetric(

backend/internal/service/ops_alerts.go

@@ -208,7 +208,11 @@ func (s *OpsService) UpdateAlertEventStatus(ctx context.Context, eventID int64,
 	if eventID <= 0 {
 		return infraerrors.BadRequest("INVALID_EVENT_ID", "invalid event id")
 	}
-	if strings.TrimSpace(status) == "" {
+	status = strings.TrimSpace(status)
+	if status == "" {
+		return infraerrors.BadRequest("INVALID_STATUS", "invalid status")
+	}
+	if status != OpsAlertStatusResolved && status != OpsAlertStatusManualResolved {
 		return infraerrors.BadRequest("INVALID_STATUS", "invalid status")
 	}
 	return s.opsRepo.UpdateAlertEventStatus(ctx, eventID, status, resolvedAt)

backend/internal/service/ops_retry.go

@@ -220,11 +220,8 @@ func (s *OpsService) RetryError(ctx context.Context, requestedByUserID int64, er
 		msg := result.ErrorMessage
 		updateErrMsg = &msg
 	}
+	// Keep legacy result_request_id empty; use upstream_request_id instead.
 	var resultRequestID *string
-	if strings.TrimSpace(result.UpstreamRequestID) != "" {
-		v := result.UpstreamRequestID
-		resultRequestID = &v
-	}
 
 	finalStatus := result.Status
 	if strings.TrimSpace(finalStatus) == "" {

backend/internal/service/ops_service.go

@@ -261,7 +261,7 @@ func (s *OpsService) ListRetryAttemptsByErrorID(ctx context.Context, errorID int
 		return nil, err
 	}
 	if s.opsRepo == nil {
-		return nil, infraerrors.NotFound("OPS_ERROR_NOT_FOUND", "ops error log not found")
+		return nil, infraerrors.ServiceUnavailable("OPS_REPO_UNAVAILABLE", "Ops repository not available")
 	}
 	if errorID <= 0 {
 		return nil, infraerrors.BadRequest("OPS_ERROR_INVALID_ID", "invalid error id")

frontend/src/i18n/locales/en.ts

@@ -155,7 +155,8 @@ export default {
     noGroupsAvailable: 'No groups available',
     unknownError: 'Unknown error occurred',
     saving: 'Saving...',
-        selectedCount: '({count} selected)',    refresh: 'Refresh',
+    selectedCount: '({count} selected)',
+    refresh: 'Refresh',
     settings: 'Settings',
     notAvailable: 'N/A',
     now: 'Now',