# Follow sync runs and logs

Read a knowledge source's status pill, its Runs tab and its Logs tab to see what each sync did, and know when to sync again or reindex.

Every sync of a knowledge source is a run, whether the schedule started it, you selected **Sync now**, or you forced a **Reindex**. A source's detail page tells the story at three depths: the status pill in its header, the **Runs** tab with one row per run, and the **Logs** tab with the engine's own log lines.

**Who can use it:** the **Admin** role or the **Knowledge Admin** grant, in a tenant licensed for knowledge, on a source's detail page in Knowledge Management in the Control Room.

## Prerequisites

- A source. Runs and logs appear once its first sync has started; a schedule of **Manual** means nothing starts one on its own.

## What a run is

**Sync now** processes only new, changed and deleted files, so once a source is fully indexed a healthy run touches few documents or none. **Reindex** re-embeds every document from scratch. Either way the source's status pill summarises the outcome:

| Status | Meaning |
| --- | --- |
| **Never synced** | No sync has ever finished, so nothing is searchable yet. |
| **Pending** | The source is created or queued and its run has not started. |
| **Syncing** or **Running** | A run is in progress. **Pause** replaces **Sync now**, the header says **Processing - stats will update automatically**, and the tiles refresh every few seconds. |
| **Synced** | The last run finished, the source holds every document the connector reported, and nothing failed. |
| **Partial** | The source holds fewer documents than the connector discovered, or the last run recorded an error or failed documents. The banner **Some documents failed** counts them. |
| **Empty** | The run finished but every document produced no chunks, so nothing is searchable. |
| **Failed** | The last run failed. The banner **Sync Failed** shows the error. |
| **Paused** | Syncs are suspended. A source that has been switched off shows **Start** instead of **Sync now**. |

An unrecognised status reads **Unknown**. An incremental run that processed a handful of documents, or none, is not partial: **Partial** is judged on what the source holds against what the connector reported, and on errors, never on how much one run touched.

## Read the Runs tab

  
**Step 1. Open the tab**

In the Control Room, open **Knowledge**, select the source's row, then **Runs**.

    ![A source's detail page on the Runs tab: the header with Sync now, Reindex and Delete, the Documents, Chunks, Last Sync and Schedule tiles, the tabs Overview, Documents, Domain graph, Access, Runs and Logs, then a card whose title counts the runs and a table with Status, Processed, Total, Errors, Started, ETA and Progress, one row per run with a status pill and Done in the Progress column.](https://docs.katonic.ai/img/knowledge/knowledge-bases/runs-and-logs/source-runs-tab.png)
*The Runs tab lists one row per sync run. Processed counts the documents that run touched, Total the documents the connector discovered, and Errors the documents that failed; Progress reads Done once a run has finished.*

    The card title counts the runs. When it reads **200 most recent runs** the table has hit its ceiling and is a sample of the history, not all of it.

  
**Step 2. Read a row**

**Status** is the run's pill: **Synced** for a completed run, **Failed** for a failed one, **Syncing** for the run in progress, which sits at the top with a live **Progress** bar. **Processed** counts the documents that run touched and **Total** the documents the connector discovered, so a small number over a large one is normal after the first sync. **Errors** counts documents that failed in that run. **Started** is when it began; **ETA** is the expected finish of a running run or the finish time of a completed one; **Progress** reads **Done** once complete.

Runs refresh on their own while a sync is active, for a minute after you start one, and while any document is still pending or processing. **No sync runs yet - trigger a sync to see run history** means no run has started.

## Read the Logs tab

  
**Step 1. Open the tab**

Select **Logs**. A dark console panel headed **Logs** opens with **Refresh** and a close control (the cross) that returns you to **Overview**. It fetches the log when you open it and does not refresh by itself.

    ![A source's detail page on the Logs tab: the header with Sync now, Reindex and Delete, the four tiles, the detail tabs, then a dark console panel headed Logs with a Refresh button and a close control, holding timestamped log lines each marked with a level.](https://docs.katonic.ai/img/knowledge/knowledge-bases/runs-and-logs/source-logs-tab.png)
*The Logs tab shows the engine's log lines for the source, each with a timestamp and a level. Select Refresh to fetch new lines; the panel does not refresh on its own.*

  
**Step 2. Read the lines**

Each line carries a timestamp, a level (**INFO**, **WARN** or **ERROR**) and the engine's message, oldest first; the panel scrolls to the newest line. While it fetches it says **Loading logs...**; with nothing recorded it says **No sync logs available yet. Trigger a sync to generate logs.** For a failed run, read the WARN and ERROR lines first.

## Act on what you see

- **Sync now**, **Pause** and **Start** are described in [Connect sources](/knowledge/knowledge-bases/connect-sources).
- **Reindex** asks you to confirm. The dialog says **Re-embeds every passage from scratch using the current embedding model. Use this after changing chunking or the embedding model, or to clear stale vectors.**, shows **Passages to re-embed**, **Est. time** and **Model**, and notes that **The source stays searchable with its existing embeddings until the reindex completes. No documents are deleted.** Select **Start reindex**.
- Reindexing is not the tool for one document: use **Re-index this document** on the **Documents** tab, described in [Browse and curate a source's documents](/knowledge/knowledge-bases/documents).

## Verify

After **Sync now**, the pill reads **Syncing** and then **Synced**, a new top row on **Runs** shows the run with **Errors** at 0 and **Progress** at **Done**, and **Refresh** on **Logs** shows the lines that run wrote.

## Troubleshooting

- **No sync runs yet - trigger a sync to see run history**: select **Sync now**, or **Start** if the source is switched off. A **Manual** schedule never runs on its own.
- **No sync logs available yet. Trigger a sync to generate logs.** on a source that has run: the log could not be fetched. Select **Refresh**; if **Couldn't load logs.** appears or the text persists, ask your administrator to check the knowledge service.
- A run stays **Syncing**: leave the tab open; the page keeps refreshing while documents are pending or processing. If the engine restarted during a run, the source comes back **Partial** and its banner says why; **Sync now** picks up the rest.
- A run shows **Errors**: open **Documents**, filter by **Failed**, read each reason and retry. If the banner reads **Sync Failed**, read the error; a changed credential is fixed under **Credentials**, then **Sync now**.
- **Partial** that does not clear: compare the **Documents** tile with **Total** on the latest run. Documents the connector reported but could not deliver (**Access denied** or **failed**) keep the source partial until they index or are deleted.

## Related

- [Browse and curate a source's documents](/knowledge/knowledge-bases/documents)
- [Connect sources](/knowledge/knowledge-bases/connect-sources)
- [Quality](/knowledge/knowledge-bases/quality)
- [Knowledge](/knowledge/overview)

Source: https://docs.katonic.ai/knowledge/knowledge-bases/runs-and-logs
