docs: document the computed outbox outcome filter

en/basic/admin-panel/computed-outbox.mdx

@@ -30,10 +30,10 @@ Each row is one task by default, and it moves through states in place as Teable
 To narrow the list:
 
 - Click the state tiles above the table to filter by state: **Waiting**, **Active**, **Delayed**, **Failed**, **Completed (retained)**, and so on. You can select several at once, and hovering a tile explains what that state means.
-- Filter by **Space**, **Base**, or **Cause**, search for a task, Base, Space, or error, or set a minimum processing time to surface slow tasks.
+- Filter by **Space**, **Base**, **Cause**, or **Outcome**, search for a task, Base, Space, or error, or set a minimum processing time to surface slow tasks.
 - Click a row to open its detail: queue timestamps, delivery lag, processing time, and, for failed deliveries, the full failure reason with the redacted SQL.
 
-<Info>**Completed** does not mean the task did work. The outcome badge says what the delivery actually did: **Processed** ran the task, **No-op** found nothing left to do, **Deferred** rescheduled it for later, and **Parked** means its scope is paused.</Info>
+<Info>**Completed** does not mean the task did work. The outcome badge says what the delivery actually did: **Processed** ran the task, **No-op** found nothing left to do, **Deferred** rescheduled it for later, and **Parked** means its scope is paused. The **Outcome** filter narrows the list by that result and shows how many deliveries each one covers, so pairing **Completed (retained)** with **Processed** leaves only the deliveries that really ran the task.</Info>
 
 A failed row is the history of one delivery attempt, not a stuck task. The durable task keeps retrying on its own, so read the **Ledger state** column to decide whether to act: **Settled** means a later retry already succeeded, and **Dead letter** links to anomaly maintenance. Settled failures are hidden by default; a banner shows how many there are and lets you reveal them.
 

zh/basic/admin-panel/computed-outbox.mdx

@@ -30,10 +30,10 @@ description: "在管理面板中监控 BullMQ 投递和计算任务积压。"
 缩小范围的方式：
 
 - 点击表格上方的状态卡片（**等待中**、**执行中**、**延迟任务**、**失败**、**已完成（保留）** 等）按状态筛选，可以多选；悬停卡片会说明该状态的含义。
-- 按 **Space**、**Base**、**触发原因** 筛选，搜索任务、Base、Space 或错误，也可以设置最短处理耗时来找出慢任务。
+- 按 **Space**、**Base**、**触发原因**、**执行结果** 筛选，搜索任务、Base、Space 或错误，也可以设置最短处理耗时来找出慢任务。
 - 点击某一行可以查看详情：入队和投递时间、投递延迟、处理耗时；失败的投递还会给出完整失败原因和脱敏后的 SQL。
 
-<Info>**已完成** 不代表任务真的做了事情。执行结果徽章说明这次投递的实际结果：**已执行** 表示执行了计算任务，**无操作** 表示已无事可做，**已推迟** 表示改到稍后重试，**已挂起** 表示所在范围被暂停。</Info>
+<Info>**已完成** 不代表任务真的做了事情。执行结果徽章说明这次投递的实际结果：**已执行** 表示执行了计算任务，**无操作** 表示已无事可做，**已推迟** 表示改到稍后重试，**已挂起** 表示所在范围被暂停。**执行结果** 筛选器可以按这些结果筛选，并显示每种结果各有多少条；把 **已完成（保留）** 和 **已执行** 一起选上，剩下的就是真正执行过计算任务的投递。</Info>
 
 失败行只是某一次投递尝试的历史记录，不代表任务卡住，持久任务会自行继续重试。是否需要介入，要看 **账本状态** 一列：**已了结** 表示后续重试已经成功，**已成死信** 可以跳转到异常任务维护。已了结的失败记录默认隐藏，横幅会显示条数并支持展开。