Overview
Every Execute dbt Project and Execute dbt Platform Job execution produces structured, browsable results. Etlworks parses and normalizes the dbt artifacts — manifest.json, run_results.json, catalog.json, and sources.json — and presents them in a dedicated dbt Results tab of the flow-execution dashboard, alongside the standard log and metrics.
The dbt Results tab
Open the execution in the dashboard and select the dbt Results tab. It shows:
- Summary cards — command, overall status, dbt and adapter versions, execution mode (local or Platform), project provenance (source, requested ref, resolved commit, subdirectory), dependency mode/status/lock/cache identity, state and retry provenance, and — for Platform runs — the remote job ID, remote status, and a link to the remote run.
- A resource table — every model, seed, test, and source with Type, Name, Status, Relation, Columns, Time (s), Details, and Lineage. Long resource names wrap rather than truncate.
- Search — filter resources by name or dbt unique ID.
- Details — per-resource details, including failure messages for failed models and tests.
The tab refreshes automatically until the execution reaches a terminal state, and renders in both light and dark themes.
Visual lineage
Click Lineage on a resource to open the lineage popup: the resource with its parents and children, each showing type, status, and relation. The graph supports fit, pan, and zoom, so wide dependency neighborhoods stay navigable.
Raw artifacts
The raw dbt artifacts of each invocation can be downloaded from the dashboard. Downloads are authenticated API calls — bearer tokens are never embedded in URLs.
Storage is bounded per account (Settings, dbt section):
| Limit | Default |
|---|---|
| Maximum Artifact File Size | 100 MiB |
| Maximum Artifacts per Run | 250 MiB |
| Artifact Storage Quota per Account | 10 GiB |
| Retention | 30 days, always preserving the latest 10 artifact-bearing executions per flow |
Root and tenant accounts have independent application-data roots and quotas. If artifact persistence fails, the execution logs a warning but a successful dbt warehouse outcome is not changed. Normalized results can remain in history after the raw files expire.
Multiple dbt invocations in one execution
When a parent execution contains several dbt invocations — nested flows, Composer branches, loops, parallel branches, retries, or a mix of dbt and non-dbt steps — the parent dashboard shows all correlated dbt invocations. A selector chooses the invocation to inspect; each is labeled with its child flow, Composer path, loop iteration, attempt, and local/Platform mode, and keeps its own raw-artifact directory.
Results in the CLI and API
The Etlworks CLI can retrieve results (dbt results), list artifacts (dbt artifacts), and download an artifact (dbt download-artifact) for a saved dbt flow. The execution APIs expose the same authenticated downloads.