dbt State in practice: what it is, how it works, and how to tune it

🌎 Global friendly sessions September 1 & 2

Ever watched a model rebuild just because someone added a comment? dbt State is how dbt tells the difference. In this session, we're getting into the details on how it actually works, with in-depth examples from the real scenarios.

What you'll learn:

• What dbt State is: a fast overview for anyone new to the concept
• How it works: the combination of checking a model’s content, the meaning of the rendered SQL and the modified times of upstream database objects lets dbt State rebuild what’s changed and skip what hasn’t.
• In-depth examples: how dbt State handles volatile SQL (like getdate()) and Jinja macros (like {{ invocation_id }}); how to configure and customize freshness checks; how time travel makes it possible to clone models based on your local code; and which configs can change the meaning of a table — and how dbt State avoids rebuilding models just because you added a tag.
• Understanding the billing model: how Daily Active Target Tables are calculated, optimization tips, and how to use the free trial.
• Installation & feedback: how to enable dbt State across dbt versions and deployment strategies, and where to share feedback

This session’s demos will focus on locally hosted dbt Core v1 deployments, but the same principles will apply if you are using dbt Core v2.0, the dbt Fusion engine, or deploying through the dbt platform.

Meet the Speakers

Joel Labes

Staff DX Advocate

dbt Labs

Reuben McCreanor

Group Product Manager

dbt Labs