Prerequisites
- A Bolt schedule you can edit.
- For an ingestion sync: your source connected to Paradime (for example, Airbyte or Fivetran) with its credentials set as Bolt schedule environment variables.
- For a Python script: the script committed to your dbt™ project.
Steps
1
Understand what a command is
A schedule is an ordered list of commands. Each command runs a step of your pipeline, and Bolt runs them top to bottom. Beyond dbt™ commands, you can add ingestion syncs (Airbyte, Fivetran, and more) and custom Python scripts to the same schedule.Order the commands to match your pipeline: sync source data first, then transform it with dbt™, then run any downstream checks or scripts.
2
Add a non-dbt command
Open the schedule and add a command. Pick the command type for the task you want to run, for example Sync Airbyte Connection, Sync Fivetran Connector, or Run Python Script.
3
Configure an ingestion sync (example)
For a Sync Fivetran Connector command, enter one or more Connector IDs (for example,
unsold_decoration). Use Add connector to sync several in parallel; the schedule continues to the next command only once all syncs finish. Then click Add command.For a Sync Airbyte Connection command, enter one or more Connection IDs (for example, conn_123), choose the Job type (Sync to run the connection, or Reset to clear its data and re-sync), then click Add command.Set the connector or connection to Manual scheduling in Fivetran or Airbyte, so syncs run only when Paradime triggers them.
4
Or run a Python script (example)
For a Run Python Script command, enter one or more commands, one per line. If you manage dependencies with Poetry, keep Without Poetry, run the script directly:
pyproject.toml and poetry.lock at the root of your dbt™ project and make poetry install the first command:Store secrets (API keys, credentials) as Bolt schedule environment variables and read them with
os.environ rather than hardcoding them.5
Order commands and add your dbt™ run
Arrange the commands so each step has what it needs: run the ingestion sync first, then your dbt™ commands to transform the fresh data, then any downstream Python checks. Save the schedule.
Trigger the schedule and open its run. You’ll know it worked when every command shows as completed in order, with the ingestion sync finishing before the dbt™ run starts. If a sync hangs, confirm the source is set to Manual scheduling and that its credentials are set as Bolt environment variables.
Next steps
Trigger runs the right way
Choose schedule types and triggers, including running a pipeline on another schedule’s completion.
Set up Turbo CI
Run and test only changed models on every pull request.
Fivetran command reference
Every field for the Sync Fivetran Connector command.
Python script reference
Poetry setup, environment variables, and usage for Run Python Script.