Skip to main content

Troubleshooting

Common issues and how to resolve them.

Authentication​

dlt runtime login fails or times out​

  • Make sure you have a valid invite code in .dlt/secrets.toml under [runtime]
  • Check your internet connection -- login requires reaching GitHub OAuth
  • Try dlt runtime logout followed by dlt runtime login to reset your session
  • See dlt runtime login for options

"Not authorized" errors​

  • Your session may have expired. Run dlt runtime login again.
  • Check that you have the correct workspace role for the action you're attempting.
  • Verify you're connected to the right workspace with dlt runtime info.

Deployment​

"No deployment found" in the Web UI​

You haven't synced your code to Runtime yet. Run:

dlt runtime deploy

See Deploy a Pipeline and the Deployment & Config page.

Changes not reflected in runs​

Make sure you've deployed your latest code before running:

dlt runtime deploy
dlt runtime launch my_pipeline.py

dlt runtime launch automatically syncs before running, but if you're using schedules, you need to deploy separately. The next scheduled run will use the latest deployment.

Runs​

Run fails immediately​

  • Check logs for error details: dlt runtime logs my_pipeline.py
  • Verify your credentials are correct in .dlt/secrets.toml. See Manage Secrets.
  • Make sure your pipeline runs locally first: python my_pipeline.py
  • Check the run detail page in the Web UI for the full stack trace.

Run stuck in "Running" state​

  • Batch jobs have a configurable maximum runtime and will be cancelled automatically when exceeded.
  • You can manually cancel: dlt runtime cancel my_pipeline.py. See dlt runtime cancel.
  • Check the run detail page in the Web UI for log output that might indicate where it's stuck.

Interactive notebook stops unexpectedly​

  • Notebooks may stop due to inactivity or resource limits.
  • Restart with dlt runtime serve my_notebook.py. See Serve a Notebook.

Schedules​

Scheduled runs not executing​

  • Verify the schedule is active: dlt runtime job info my_pipeline.py
  • Check the Jobs page for the Next Run column
  • Make sure you have a valid deployment -- scheduled runs use the latest deployed code

Wrong schedule time​

  • Cron expressions use UTC time
  • Update the schedule: dlt runtime schedule my_pipeline.py "0 9 * * *"
  • See Schedule a Pipeline for cron pattern examples

Known limitations​

  • Batch jobs have a configurable maximum runtime and will be cancelled automatically when exceeded
  • For more details, see the Runtime limitations in the dlt docs

Getting help​