> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cloosphere.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Retention

> Set automatic cleanup retention periods for log/trace data and check the background task queue status.

<Info>Admin › Settings › Data Retention</Info>

Set the **retention period** for accumulating log data such as usage, audit, and trace, and automatically clean up data past the period.

* Use this to manage storage and comply with data retention policies.

<Frame caption="Admin > Settings > Data Retention — automatic/manual cleanup and task queue">
  <img src="https://mintcdn.com/cloocus/3CYTE_uN2N3A4joq/images/admin/settings-data-retention.png?fit=max&auto=format&n=3CYTE_uN2N3A4joq&q=85&s=63dfa873411ccd2a6108c5664fee8f14" alt="Data Retention settings" width="2880" height="1800" data-path="images/admin/settings-data-retention.png" />
</Frame>

***

## Automatic Cleanup

When **Enable Auto Cleanup** (off by default) is on, data past the configured retention period is deleted automatically every day.

* Turning the toggle on also reveals a **Cleanup Time** dropdown. Pick the hour the cleanup job runs; the default is `03:00`.

<Warning>
  Setting a retention period to **`0` means permanent retention** (no automatic deletion). To clean up, specify a number of days of 1 or more. **Every type defaults to `0`**, so until you enter values, turning auto cleanup on deletes nothing.
</Warning>

### Retention Targets

Set the retention period (days) individually for each data type.

* Under each type name, the **current record count and storage size** are shown, so you can tell what to trim first.

| Data                | Description                                           |
| ------------------- | ----------------------------------------------------- |
| **Usage logs**      | Token/request usage records                           |
| **Audit logs**      | Admin/user activity audit records                     |
| **Guardrail logs**  | Guardrail block/pass records                          |
| **Traces**          | Step-by-step records of how AI requests are processed |
| **Trace analytics** | Aggregated/summarized analytics of traces             |
| **Auto evaluation** | Automatic evaluation results                          |

***

## Manual Cleanup

Clicking **Run Cleanup Now** saves the current settings and **immediately** deletes data past the retention period.

<Note>
  If every retention period is `0`, nothing runs and the message "Set retention period for at least one log type before cleanup" is shown.
</Note>

***

## Task Queue

Check the status of the work queue that processes heavy tasks (like file processing) in the background.

* The queue backend is shown next to the description — **Redis** when connected to Redis, otherwise **In-process** (handled inside the server process).

| Item           | Description                             |
| -------------- | --------------------------------------- |
| **Queued**     | Number of tasks waiting to be processed |
| **Processing** | Number of tasks currently running       |
| **Workers**    | Number of workers processing tasks      |

### Consumers & Stuck Messages

Expanding `▶ Consumers & Stuck Messages` shows the **Consumers (N)** list and the **Oldest pending messages (N)** list, along with two buttons.

| Button                  | Action                                                                                           |
| ----------------------- | ------------------------------------------------------------------------------------------------ |
| **Cleanup Zombies**     | Removes only consumers idle for more than an hour that have no pending messages. No data is lost |
| **Reclaim Stuck (>5m)** | Force-acknowledges pending messages that have been stuck for more than 5 minutes                 |

<Warning>
  **Reclaiming drops the work instead of running it.** Use it only when the queue is jammed, and check the list first to see which tasks disappear.
</Warning>