fix: 补充缺失的组级和账户级运维告警指标

新增以下运维告警指标类型:
- group_available_accounts: 组内可用账户数
- group_available_ratio: 组内可用账户比例
- group_rate_limit_ratio: 组内限速账户比例
- account_rate_limited_count: 限速账户数
- account_error_count: 错误账户数
- account_error_ratio: 错误账户比例
- overload_account_count: 过载账户数

包含比例和计数类指标的评估逻辑，并注册新的百分比类指标用于阈值校验。

backend/internal/handler/admin/ops_alerts_handler.go

@@ -23,6 +23,13 @@ var validOpsAlertMetricTypes = []string{
 	"cpu_usage_percent",
 	"memory_usage_percent",
 	"concurrency_queue_depth",
+	"group_available_accounts",
+	"group_available_ratio",
+	"group_rate_limit_ratio",
+	"account_rate_limited_count",
+	"account_error_count",
+	"account_error_ratio",
+	"overload_account_count",
 }
 
 var validOpsAlertMetricTypeSet = func() map[string]struct{} {
@@ -82,7 +89,10 @@ func isPercentOrRateMetric(metricType string) bool {
 		"error_rate",
 		"upstream_error_rate",
 		"cpu_usage_percent",
-		"memory_usage_percent":
+		"memory_usage_percent",
+		"group_available_ratio",
+		"group_rate_limit_ratio",
+		"account_error_ratio":
 		return true
 	default:
 		return false