.NET 6中MongoDB FindAsync().ToList()性能不及文件读取的优化方案

我在MongoDB中存储了约600份包含嵌套列表的城市酒店搜索响应文档，已为RequestId字段创建索引。使用MongoDB Compass执行查询{"RequestId": "1232423"}仅需130毫秒，但在.NET 6中执行以下代码时，耗时高达2500-3000毫秒：

_search2.FindAsync(e => true && e.RequestId == RequestId).Result.ToList();

甚至将相同JSON内容存储在文件系统中，读取并反序列化仅需约1000毫秒，迁移至MongoDB后的性能反而未达预期。以下是该查询的executionStats：

{
  "explainVersion": "1",
  "queryPlanner": {
    "namespace": "test.CitySearch",
    "indexFilterSet": false,
    "parsedQuery": { "RequestId": { "$eq": "1232423" } },
    "maxIndexedOrSolutionsReached": false,
    "maxIndexedAndSolutionsReached": false,
    "maxScansToExplodeReached": false,
    "winningPlan": { "stage": "EOF" },
    "rejectedPlans": []
  },
  "executionStats": {
    "executionSuccess": true,
    "nReturned": 0,
    "executionTimeMillis": 0,
    "totalKeysExamined": 0,
    "totalDocsExamined": 0,
    "executionStages": {
      "stage": "EOF",
      "nReturned": 0,
      "executionTimeMillisEstimate": 0,
      "works": 1,
      "advanced": 0,
      "needTime": 0,
      "needYield": 0,
      "saveState": 0,
      "restoreState": 0,
      "isEOF": 1
    }
  },
  "command": {
    "find": "CitySearch",
    "filter": { "RequestId": "1232423" },
    "$db": "test"
  },
  "serverInfo": {
    "host": "LAPTOP-LH7H8EFV",
    "port": 27017,
    "version": "7.0.5",
    "gitVersion": "7809d71e84e314b497f282ea8aa06d7ded3eb205"
  },
  "serverParameters": {
    "internalQueryFacetBufferSizeBytes": 104857600,
    "internalQueryFacetMaxOutputDocSizeBytes": 104857600,
    "internalLookupStageIntermediateDocumentMaxSizeBytes": 104857600,
    "internalDocumentSourceGroupMaxMemoryBytes": 104857600,
    "internalQueryMaxBlockingSortMemoryUsageBytes": 104857600,
    "internalQueryProhibitBlockingMergeOnMongoS": 0,
    "internalQueryMaxAddToSetBytes": 104857600,
    "internalDocumentSourceSetWindowFieldsMaxMemoryBytes": 104857600,
    "internalQueryFrameworkControl": "trySbeRestricted"
  },
  "ok": 1
}

优化步骤 #

1. 清理冗余查询条件 #

LINQ查询中的true &&属于无意义冗余，虽然驱动可能自动优化，但建议直接简化为：

_search2.FindAsync(e => e.RequestId == RequestId).Result.ToList();

避免驱动解析时的额外开销。

2. 替换同步阻塞为异步await #

使用.Result会强制阻塞线程，引发不必要的上下文切换，改为异步模式：

await _search2.FindAsync(e => e.RequestId == RequestId).ToListAsync();

异步模式能更高效利用线程资源，减少等待耗时。

3. 验证索引有效性与数据一致性 #

从executionStats的nReturned:0来看，本次查询未匹配到文档，但Compass能查到结果，需确认：

• .NET代码中传入的RequestId值与Compass完全一致（注意大小写、空格、特殊字符）
• 索引是否真实存在：在MongoDB shell执行db.CitySearch.getIndexes()查看RequestId索引状态
• 文档中RequestId的类型是否统一：避免字符串与数字类型混用导致索引无法匹配

4. 优化序列化/反序列化性能 #

MongoDB .NET驱动的序列化开销可能是主要瓶颈，可做以下调整：

• 用[BsonIgnoreExtraElements]标记实体类，跳过文档中不存在的字段解析
• 自定义序列化配置，针对嵌套列表启用更高效的序列化逻辑
• 开启驱动性能日志，定位序列化阶段耗时：

  var settings = MongoClientSettings.FromConnectionString("你的连接字符串");
  settings.ClusterConfigurator = cb => cb.Subscribe<CommandStartedEvent>(e =>
  {
      Console.WriteLine($"执行命令: {e.Command.ToJson()}");
  });
  var client = new MongoClient(settings);
  

5. 调整MongoDB驱动连接配置 #

• 增大连接池大小：默认连接池可能过小，导致请求等待可用连接

  var settings = MongoClientSettings.FromConnectionString("你的连接字符串");
  settings.MaxConnectionPoolSize = 100; // 根据业务量调整
  var client = new MongoClient(settings);
  

• 缩短超时时间并启用直接连接（本地测试场景）：避免DNS解析或路由开销

  settings.ConnectTimeout = TimeSpan.FromSeconds(5);
  settings.ServerSelectionTimeout = TimeSpan.FromSeconds(5);
  

6. 拆分步骤定位瓶颈 #

拆分查询与序列化步骤，明确耗时阶段：

var stopwatch = Stopwatch.StartNew();
var cursor = await _search2.FindAsync(e => e.RequestId == RequestId);
stopwatch.Stop();
Console.WriteLine($"FindAsync阶段耗时: {stopwatch.ElapsedMilliseconds}ms");

stopwatch.Restart();
var list = await cursor.ToListAsync();
stopwatch.Stop();
Console.WriteLine($"ToListAsync序列化阶段耗时: {stopwatch.ElapsedMilliseconds}ms");

内容的提问来源于stack exchange，提问作者user23194681