# Data source syncs

Every time a connector collects data, Mesh records it. **Data source syncs** is
where you read those records: one row per run, newest first.

Use it to answer three questions:

| Question | Where to look |
| --- | --- |
| Did last night's collection work? | The **Status** column |
| How much did it bring in? | The **Items** column |
| Why did it fail, and what do I fix? | Open the row, then [Troubleshoot a sync](./troubleshoot-a-sync.mdx) |

## Where to find it

1. Select **Settings**.
2. Select **Data source syncs**.

{/* ---------------------------------------------------------------------------
     SYNTHETIC SCREENSHOT DATA — NOT A RAW CAPTURE, AND NOT CUSTOMER DATA.

     The screenshots below show the REAL product UI (layout, columns, badges,
     tooltips and every product string are exactly as they ship) populated with
     EXAMPLE data. Before each capture, tenant-identifying values were replaced:

       - connector names  -> Okta, Microsoft Entra ID, Workday, ServiceNow CMDB,
                             Salesforce, Google Workspace, Active Directory,
                             TriNet, Jira
       - every UUID       -> deterministic example ids, generated from a hash so
                             one connector keeps one id across all images
       - account name     -> an example name
       - stat-card totals -> example figures, consistent with the product's own
                             success-rate formula

     No real connector name, tenant identifier, account name or configuration
     UUID appears in any image. Verified per capture by scanning the rendered
     page, and the ids are reproducible from the generator.

     Full substitution record: internal-notes/screenshot-provenance/data-source-syncs.md
     (deliberately outside docs/ and static/, so it is never published).
--------------------------------------------------------------------------- */}

![The Data source syncs tab: five summary cards above a table of runs, one row per run](/img/admin-guide/data-source-syncs/01-overview.png)

## One row per run

A single collection has a start and a finish. Data source syncs joins them into
**one row**, so the count of rows is the count of runs — not double it.

A run that has started but not finished shows as **In progress**. When it
finishes, the same row updates with the outcome and the duration.

## What it covers

- **Every connector type.** Whatever you have connected appears here.
- **Scheduled and manual runs.** A run you triggered yourself is recorded the
  same way as one that ran on its schedule, and the **Trigger** column tells
  you which it was.
- **The last 90 days**, selectable in the time-range filter.

> **note**
The **Items** column is what that one run collected — not a running total for
the connector. A connector that collects nothing new on an incremental run is
working correctly, and shows a low number.

## Two screens, two questions

Data source syncs sits alongside the connector list, and they answer different
questions. Both are useful; use the one that matches what you are asking.

| Screen | Answers |
| --- | --- |
| **Connectors** list | Is this connector healthy **right now**? |
| **Data source syncs** | What happened on **each individual run**? |

A connector can be healthy today and still have a failed run last week. That is
normal, and the two screens showing different things is not a contradiction —
they are reporting different time frames.

## Next

| Page | What it covers |
| --- | --- |
| **[Read a sync result](./read-a-sync-result.mdx)** | Every column, every status, and what the item counts mean |
| **[Troubleshoot a sync](./troubleshoot-a-sync.mdx)** | A failed run, a partial run, and a run that never finishes |