Bolt Logs Tool
The Bolt Logs Tool allows DinoAI to access and analyze execution logs from your Bolt scheduled jobs, bringing run history, error details, and performance data directly into your development workflow.
This tool connects DinoAI to your Bolt scheduler, enabling it to inspect dbt run outputs, diagnose failures, and help you resolve issues based on real log data β without leaving the Code IDE.
Capabilities
Fetch Run Logs
Retrieve execution logs for any Bolt schedule run by ID or schedule name
Diagnose Failures
Identify the root cause of failed dbt models or tests from log output
Analyse Run History
Review recent runs for a schedule to spot patterns or regressions
Surface Error Details
Extract and explain compiler errors, runtime exceptions, and test failures
Compare Runs
Contrast a passing run against a failing one to isolate what changed
Using the Bolt Logs Tool
Open DinoAI in the right panel of the Code IDE
Reference the Bolt schedule or run you want to inspect (e.g.,
"the last run of the nightly schedule"or a specific run ID)Add your prompt describing what you want DinoAI to investigate
Grant permission when DinoAI asks to access Bolt logs
Review DinoAI's analysis and follow the suggested fixes
Example Use Cases
Diagnosing a Failed Run
Prompt
Result: DinoAI fetches the latest run logs for the schedule, identifies the failing models or tests, extracts the relevant error messages, and suggests the most likely fixes β pointing you directly to the affected files.
Investigating a Slow Run
Prompt
Result: DinoAI reviews the run timing across models, highlights the slowest nodes, and suggests optimisations such as incremental model adjustments, test pruning, or upstream source query improvements.
Comparing a Passing and Failing Run
Prompt
Result: DinoAI retrieves both runs, diffs the outputs, identifies which models started failing and when, and correlates the failure with recent code or schema changes.
Working with Other Tools
The Bolt Logs Tool works well alongside DinoAI's other capabilities for a complete debugging workflow:
Combine with the File System Tool to apply fixes to the models identified as failing directly from the log analysis
Combine with the Terminal Tool to re-run specific dbt models or tests after a fix has been applied
Use alongside the SQL Execution Tool to validate model logic against live warehouse data after diagnosing a data quality failure
Pair with the Snowflake or BigQuery Query Analysis Tools to investigate upstream performance issues that caused a Bolt job to time out
Best Practices
Be specific about which schedule or run β Reference the exact schedule name or run ID in your prompt so DinoAI retrieves the correct logs without ambiguity
Provide context about what you expected β Telling DinoAI what the job should produce helps it identify deviations faster
Use it iteratively β After applying a fix, ask DinoAI to check the next run's logs to confirm the issue is resolved
Review full error chains β dbt failures often cascade; ask DinoAI to trace the root model failure rather than fixing every downstream symptom individually
Last updated
Was this helpful?