Browse sources and record pages

Understand run-and-step grouping, loaded counts, and the Unsorted bucket.

The Data overview groups records into sources. A source can identify one workflow run and one producing step, so two steps in the same run can appear separately. A connector-ingested source can represent a run without that step subdivision.

Open the appropriate source

Use the source's name and record types to choose the output you want, then open its record list. Returning from an individual record preserves the source context when you opened the record from that list.

A cold link can initially show a generic source name while its source summary is not loaded. Use the record's metadata and provenance to confirm its origin rather than relying only on a temporary label.

Read counts carefully

The number shown in a browsing view can describe the records loaded or filtered so far. It should not be treated as a complete inventory until paging is finished. A table can have more records available beyond the current window.

Type filters narrow source membership and record results. A source's displayed type information helps locate it; it does not imply every row in that source has identical content. Open the record to inspect its actual type and payload.

Unsorted records

The Unsorted category holds records without a normal source attribution. It helps keep those records discoverable without inventing a workflow provenance. It cannot be deleted as though it were an ordinary source run.

When investigating missing records, clear filters, load more pages, and check whether the output belongs to a different step or run. A successful workflow status alone does not establish that a particular step produced the type or number of records you expected.

For interpreting record payloads, see record content. For the effect of removing a source group, see source deletion.