v1.7 upgrade guide: updates to model governance features

[jtcohen6]

Contributions

• I have read the contribution docs, and understand what's expected of me.

Link to the page on docs.getdbt.com requiring updates

https://docs.getdbt.com/guides/migration/versions/upgrading-to-v1.7

What part(s) of the page would you like to see updated?

Add a section for updates related to "Model governance" features (introduced in v1.5 + refined in v1.6)

Interaction between model contracts & versions

Breaking change detection for models with enforced contracts (#4000): When dbt detects a breaking change to a model with an enforced contract during state comparison, dbt will now raise an error for versioned models and a warning for models that are not versioned. (In previous versions, this was always an error.) The addition of a version communicates that this model is a stable API, and any breaking changes will be accompanied by a version bump. Breaking changes include: removing a column, renaming a column, changing its data type, or changing/removing an enforced constraint.

Flexibility around setting model `access

Set access as a config (#4075): It's now possible to set a model's access within config blocks: within the model's file, or within dbt_project.yml for an entire subfolder at once. We strongly advise you to be thoughtful about which models you are marking public! In a "Mesh" pattern, this means that other teams/projects could start referencing that model and using it as a critical dependency.

Model contracts - data type ergonomics

Type aliasing for model contracts (#4241): dbt will use each adapter's built-in type aliasing for user-provided data types—meaning you can now write string always, and dbt will translate to text on Postgres/Redshift. This is on by default, and it's possible to opt out.

Raise warning for numeric types (#4251): This led to a lot of stubbed toes, when folks put numeric in their model contracts — without realizing that Snowflake interprets it as numeric(38,0), and will coerce/round decimals accordingly. dbt will now raise a warning if it sees a numeric type without specified precision/scale.

Additional information

No response