Skip to main content
When a production run fails, Bolt’s logs and artifacts help you find the cause fast. This guide walks from spotting a failure to fixing it.
Prerequisites
  • Access to Bolt in your Paradime workspace.
  • At least one configured Bolt schedule.
  • Basic dbt™ and SQL knowledge.
Estimated time: 15 minutes.

Steps

1

Find the failed run

Two ways: a Slack or email notification (recommended for production), or the Status column on the Bolt home screen, where failures show an Error indicator.
2

Open the run logs

Select the failed schedule, open Run history, select the failed run, scroll to Logs and artifacts, and select the command that failed (for example dbt run). Bolt provides three log types:
Console logs are usually the most useful: they show the error, warnings, and the compiled SQL that failed in production.
3

Read the summary, then the console logs

Start with the summary for a quick read, then jump to the error in the console logs and open the compiled SQL link.
Summary logs example
4

Test and fix

Copy the compiled SQL from the console logs and test it directly in your warehouse or in the Code IDE scratchpad. Fix the common causes: invalid column names, missing model references, SQL syntax errors, or data type mismatches. Then commit the fix and let the schedule re-run.
You’ve located the failing command, identified the error in the console logs, and confirmed a fix by running the compiled SQL against your warehouse.

Next steps

Set up notifications

Get alerted the moment a run fails.

Run details reference

Logs, artifacts, and retrying a run.