Investigate slow runs

Outcome

You see which models or tests took the longest in a run and can explore execution timelines interactively.

When to use this

Surface	Use when
CLI	query-executions for sorted durations or CSV/JSON export
MCP	Skip unless an agent will run many execution queries
Web	Primary surface—execution views, timelines, and bottlenecks

Steps

1. Confirm artifacts with Check run health.
2. Start the web UI and open the URL printed in the terminal.
3. Use execution and timeline views to find slow nodes; use CLI query-executions for scripted top-N lists.

Example

npx @dbt-tools/web --dbt-target ./target

dbt-tools query-executions --dbt-target ./target --sort duration --limit 20 --json

Next

• Investigation tour
• Web getting started
• CLI README — query-executions filters