diff --git a/frontend/src/views/DocList.vue b/frontend/src/views/DocList.vue
index eb231c5..2448e46 100644
--- a/frontend/src/views/DocList.vue
+++ b/frontend/src/views/DocList.vue
@@ -1,5 +1,5 @@
 <template>
-  <t-card title="文档列表" :bordered="false">
+  <t-card class="doc-list-card" title="文档列表" :bordered="false">
     <div class="doc-layout">
       <!-- 左侧：目录树面板 -->
       <aside class="doc-tree-panel">
@@ -522,15 +522,19 @@ loadData()
 .selected-count { font-size: 12px; font-weight: 600; color: var(--td-text-color-primary); }
 
 /* 左右分栏布局 */
-.doc-layout { display: flex; gap: 16px; align-items: flex-start; }
+/* 卡片撑满内容区，实现左右独立滚动 */
+.doc-list-card { height: 100%; display: flex; flex-direction: column; }
+.doc-list-card :deep(.t-card__body) { flex: 1; min-height: 0; overflow: hidden; }
+
+.doc-layout { display: flex; gap: 16px; align-items: stretch; height: 100%; }
 .doc-tree-panel {
   width: 240px;
   flex-shrink: 0;
-  min-height: 420px;
+  overflow-y: auto;
   padding-right: 12px;
   border-right: 1px solid var(--color-border);
 }
-.doc-content { flex: 1; min-width: 0; }
+.doc-content { flex: 1; min-width: 0; overflow-y: auto; }
 
 .tree-panel-header { display: flex; justify-content: space-between; align-items: center; margin-bottom: 8px; }
 .tree-panel-title { font-size: 13px; font-weight: 600; }
diff --git a/src/main/java/com/wok/supportbot/config/AsyncExecutorConfig.java b/src/main/java/com/wok/supportbot/config/AsyncExecutorConfig.java
index 8781235..0c5c68d 100644
--- a/src/main/java/com/wok/supportbot/config/AsyncExecutorConfig.java
+++ b/src/main/java/com/wok/supportbot/config/AsyncExecutorConfig.java
@@ -33,4 +33,22 @@ public class AsyncExecutorConfig {
         executor.initialize();
         return executor;
     }
+
+    /**
+     * 文档异步处理线程池（分块 → 关键词提取 → 向量化）。
+     * 有界线程池 + CallerRunsPolicy：避免文件夹上传大量文件时
+     * 默认 {@code SimpleAsyncTaskExecutor} 每任务新建线程导致的线程爆炸，
+     * 同时限制并发向量化调用，避免打爆大模型限流。
+     */
+    @Bean("documentExecutor")
+    public ThreadPoolTaskExecutor documentExecutor() {
+        ThreadPoolTaskExecutor executor = new ThreadPoolTaskExecutor();
+        executor.setCorePoolSize(2);
+        executor.setMaxPoolSize(4);
+        executor.setQueueCapacity(500);
+        executor.setThreadNamePrefix("doc-process-");
+        executor.setRejectedExecutionHandler(new ThreadPoolExecutor.CallerRunsPolicy());
+        executor.initialize();
+        return executor;
+    }
 }
diff --git a/src/main/java/com/wok/supportbot/service/DocumentProcessingService.java b/src/main/java/com/wok/supportbot/service/DocumentProcessingService.java
index 4c8320a..fcb9a36 100644
--- a/src/main/java/com/wok/supportbot/service/DocumentProcessingService.java
+++ b/src/main/java/com/wok/supportbot/service/DocumentProcessingService.java
@@ -11,8 +11,6 @@ import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.jdbc.core.JdbcTemplate;
 import org.springframework.scheduling.annotation.Async;
 import org.springframework.stereotype.Service;
-import org.springframework.transaction.annotation.Propagation;
-import org.springframework.transaction.annotation.Transactional;
 
 import java.util.HashMap;
 import java.util.List;
@@ -43,7 +41,8 @@ public class DocumentProcessingService {
 
     /**
      * 异步处理文档：分块 → 关键词提取 → 向量化 → 更新状态
-     * 使用 REQUIRES_NEW 在独立事务中执行，避免与主事务冲突
+     * 不加跨方法事务：AI 关键词提取与向量化属于慢速网络调用，期间不占用数据库连接，
+     * 避免文件夹批量上传时多个异步任务把连接池占满导致连接超时。
      *
      * @param docId     文档ID
      * @param documents 已解析的原始文档列表
@@ -54,8 +53,7 @@ public class DocumentProcessingService {
      * @param chunkSize 分块大小（可选，覆盖全局配置）
      * @param overlap   重叠大小（可选，覆盖全局配置）
      */
-    @Async
-    @Transactional(propagation = Propagation.REQUIRES_NEW, rollbackFor = Exception.class)
+    @Async("documentExecutor")
     public void processDocumentAsync(Long docId, List<Document> documents, String sourceName,
                                      String title, Long categoryId, List<String> tags,
                                      Integer chunkSize, Integer overlap) {
@@ -117,8 +115,7 @@ public class DocumentProcessingService {
      * @param docId 文档ID
      * @param documents 解析后的文档列表
      */
-    @Async
-    @Transactional(propagation = Propagation.REQUIRES_NEW, rollbackFor = Exception.class)
+    @Async("documentExecutor")
     public void reprocessDocumentAsync(Long docId, List<Document> documents) {
         // 等待主事务提交，确保文档记录可见
         KnowledgeDocument doc = waitForDocument(docId);