Firebase事务执行缓慢如何优化？附相关代码求助

我有一个函数，每当用户创建新的活动参与文档时触发执行。由于部分场景下新参与文档会快速连续创建，为避免竞态条件我使用了事务，但该函数通常需要40-120秒才能完成运行。以下是相关代码：

export const esLatDailyAggr = async (
  uid: admin.firestore.DocumentReference,
  eventTypeId: admin.firestore.DocumentReference,
  timestamp: Date,
  clientId: admin.firestore.DocumentReference,
  year: number,
  month: number,
  day: number) => {
  const startTime = Date.now();
  console.log("Starting to update daily aggr data.");
  const docId = `${uid.id}-${year}-${month}-${day}`; // create composite key

  const statsRefAggr = admin.firestore().collection("dailyAggregateData");
  const queryAggr = statsRefAggr.doc(docId);
  await admin.firestore().runTransaction(async (transaction) => {
    const queryAggrSnapshot = await transaction.get(queryAggr);

    if (!queryAggrSnapshot.exists) { // check if document exists
      console.log(
        docId, " Daily doc not found. Creating new one by new Participation");
      // create new document with composite key
      transaction.set(queryAggr, {
        uid: uid,
        year: year,
        month: month,
        day: day,
        NoOfParticipations: 1,
        ActiveClientsId: [clientId],
        ActiveEventTypeId: [eventTypeId],
        timestamp: timestamp,
      });
    } else {
      // update existing document
      transaction.update(queryAggr, {
        NoOfParticipations: admin.firestore.FieldValue.increment(1),
        ActiveClientsId: admin.firestore.FieldValue.arrayUnion(clientId),
        ActiveEventTypeId: admin.firestore.FieldValue.arrayUnion(eventTypeId),
      });
      console.log(docId, " Daily aggr data updated by new Participation.");
    }
  });
  const endTime = Date.now();
  console.log("Daily Aggr Data function finished. Running Time: ",
  endTime - startTime, "ms");
};

优化方案 #

• 用原子操作+合并写入替代事务
  你代码里用到的increment和arrayUnion本身就是Firestore的原子操作，完全不需要事务来保障竞态安全。事务的锁机制和重试逻辑反而会在高并发场景下拖慢执行速度。直接使用set方法的merge: true参数，就能实现“不存在则创建，存在则更新”的逻辑，性能提升非常明显。

  修改后的代码：

  export const esLatDailyAggr = async (
    uid: admin.firestore.DocumentReference,
    eventTypeId: admin.firestore.DocumentReference,
    timestamp: Date,
    clientId: admin.firestore.DocumentReference,
    year: number,
    month: number,
    day: number) => {
    const startTime = Date.now();
    console.log("Starting to update daily aggr data.");
    const docId = `${uid.id}-${year}-${month}-${day}`;
  
    const queryAggr = admin.firestore().collection("dailyAggregateData").doc(docId);
  
    // 合并写入：文档不存在则创建，存在则仅更新指定字段
    await queryAggr.set({
      uid: uid,
      year: year,
      month: month,
      day: day,
      NoOfParticipations: admin.firestore.FieldValue.increment(1),
      ActiveClientsId: admin.firestore.FieldValue.arrayUnion(clientId),
      ActiveEventTypeId: admin.firestore.FieldValue.arrayUnion(eventTypeId),
      timestamp: timestamp,
    }, { merge: true });
  
    const endTime = Date.now();
    console.log("Daily Aggr Data function finished. Running Time: ", endTime - startTime, "ms");
  };
  

• 精简日志输出
  高并发场景下频繁打印日志会占用额外资源，建议只保留函数启动、结束的关键日志，或者仅在异常时打印，减少I/O开销。

• 优化函数运行环境（针对Cloud Function）
  如果这是Google Cloud Function，慢启动可能是原因之一：

  • 升级函数实例配置（比如增加内存配额），提升CPU和网络性能
  • 启用函数预热机制，避免冷启动延迟
  • 清理不必要的依赖包，缩小部署包体积，加快初始化速度

内容的提问来源于stack exchange，提问作者Attila