Deploy a Pipeline
This guide walks through deploying your own dlt pipeline to Runtime and verifying it runs successfully.
If you haven't deployed anything yet, start with the Quickstart using the starter pack.
Prerequisites
- You're logged in (
dlt runtime login). Seedlt runtime login. - Your pipeline runs successfully locally (e.g.,
python my_pipeline.py) - Your credentials are configured in
.dlt/secrets.toml. See Manage Secrets.
Deploy and run in one step
The simplest way to deploy and run is dlt runtime launch:
dlt runtime launch my_pipeline.py
This:
- Syncs your code (deployment) and secrets (configuration) to Runtime
- Creates a job for
my_pipeline.pyif one doesn't exist - Starts a new run
Add --follow to stream logs until the run completes:
dlt runtime launch my_pipeline.py --follow
Deploy without running
To sync your code and configuration without triggering a run, use dlt runtime deploy:
dlt runtime deploy
This updates both the deployment (code) and configuration (secrets) to their latest versions. You can then start runs separately via the Jobs page or with dlt runtime job-run create.
Sync code or configuration independently
To sync only one or the other:
dlt runtime deployment sync # code only
dlt runtime configuration sync # secrets only
See dlt runtime deployment sync and dlt runtime configuration sync.
Verify the run
From the CLI
Check your workspace status:
dlt runtime info
View logs for the latest run:
dlt runtime logs my_pipeline.py
See dlt runtime logs for options like viewing a specific run number.
From the Web UI
Open the dashboard:
dlt runtime dashboard
- The Dashboard shows the run in the Job Runs panel
- The Jobs page shows the job with its run history
- Click the run to see detailed logs and pipeline run information
What gets deployed
- Deployment: All Python files and supporting modules in your workspace directory
- Configuration: The
.dlt/directory contents (secrets, config files, profile-specific settings)
Both are versioned independently. See the Deployment & Config page to inspect current versions.
Next steps
- Schedule a Pipeline for automatic recurring runs
- Monitor and Debug to track run health and diagnose failures