The run log holds one row for every run of a table: its outcome, the rows it handled, and any error. Open it when a run fails and you need the reason, or to check a table's runs.
The log lives in the table's popup. Deploy and operate a table covers deploying the table and its other actions.
The log lists the table's runs. Its columns are:
| Column | What it shows |
|---|---|
| Log ID | The identifier of the run. |
| Status | The state of the run. |
| Rows | The number of rows the run processed. |
| Error | The run's error. The cell previews the first line; click it to open the full text. |
| User | Who started the run. |
| Start | When the run started. |
| Finish | When the run finished. |
| Duration | How long the run took. While a run is still going, the cell counts the time so far; the run's own duration arrives when it finishes. |
| Snap Date | The date the run is recorded for. |
Each row's actions sit in a column pinned to the right edge of the log.
Two things to know about failures:
These windows are shared across table popups; Shared elements in table popups gathers them with the other common pieces.
Open a column's menu and filter by that column: tick values from a list, pick dates, or set a range for numbers. To search across the log instead, type in the search box.
It matches rows where that column is blank. Pick it to find runs with no value in the cell.
Yes. Every column sorts from its own menu, and that menu also carries the column's filters.
The run has not finished, so there is no final duration yet; the cell counts the time it has run so far. The final duration appears when the run ends.
It previews the error's first line so rows stay readable. Click the preview to open the full text.
Open the full error and click Copy Error. The label turns Copied once the text is on the clipboard, then changes back. A long message scrolls inside the window.
When the server sends its own message for a failure, that message is shown first. If there is none, you see the error's own text, and a general fallback message when even that is missing.
For a run that loaded through an ANY API connector, its loader's report appears above the raw stack. Start there when the stack alone does not explain the failure.