Skip to content

Models Page

Browse and run dbt models from the web interface.

Overview

The Models page (/models) displays all dbt models in your project with their current status. Use it to check which models have run successfully and to trigger individual model runs from the browser.

Navigate to Models in the top navigation bar.

Models Table

The table shows all models from your dbt project:

Column Description
Model Model name (e.g., stg_customers, fct_orders)
Schema Database schema the model materializes into (e.g., raw, staging, marts)
Type Materialization type — table, view, or incremental
Rows Number of rows in the materialized model
Status Current run status (see Status Indicators below)
Last Run Timestamp of the most recent dbt run for this model
Actions Run button to trigger a dbt run for the model

Status Indicators

  • Success (green) — model ran successfully on its last execution
  • Error (red) — model failed on its last run
  • Running (yellow, pulsing) — model is currently being executed via the UI
  • Skipped (gray) — model was skipped in the last run
  • Not Run (gray) — model has not been run yet

Running a Model

  1. Find the model in the table.
  2. Click the Run action button in its row.
  3. A toast notification confirms the run has started.
  4. The status updates in real time via WebSocket as the run progresses.

The table provides two run actions:

  • Run — runs just the selected model (dbt run --select model_name)
  • Run+ — runs the model and all its downstream dependents (dbt run --select model_name+), keeping dependent models in sync

DuckDB write lock

dbt runs acquire the DuckDB write lock. If a sync is already running, the model run waits for the lock to be released. Only one write operation can run at a time.

Refresh

Click the refresh icon (top-right of the table header) to reload the models list from the dbt manifest. This picks up changes to your dbt project without a full page refresh.

How it works
  • List models: GET /api/dbt/models returns all models from the dbt manifest with status metadata.
  • Run a model: POST /api/dbt/models/{model_name}/run starts a background task that acquires a dbt lock, runs the model via subprocess, and refreshes the Metabase connection on success. Runs have a 5-minute timeout.
  • WebSocket events: The page receives dbt_run_started, dbt_run_progress, dbt_run_completed, and dbt_run_failed events for real-time updates and toast notifications.

Troubleshooting

Models table shows "Loading dbt models..."
dbt has not been run yet or the manifest file is missing. Run dbt run from the CLI to generate the manifest, or trigger a sync (which runs dbt automatically).
Model run fails with lock error
Another write operation (sync or dbt run) is in progress. Wait for it to complete and try again. The toast notification shows which operation holds the lock.
Model run times out
Model runs have a 5-minute timeout. For large models, run from the CLI with dbt run --select model_name which has no timeout limit.