Modelling correctness · rule catalog-model-not-built
dbt models that do not exist in the warehouse
A Metacenta review checks this under the rule Declared models exist in the warehouse. Everything below applies whether or not you ever commission one.
What this rule checks
This rule flags models declared in the project that are missing from the catalog. It runs only when the catalog covers at least 80% of the project's models. It declines if the adapter reported any introspection error.
Why it matters
Anything selecting from the model in that environment fails. Its absence is invisible to any run that never selected it, so nobody learns until a query breaks.
How to fix it
Build or retire the models missing from the warehouse. Build the model, or retire it from the project if it is no longer wanted.
When it is fine to leave
A model that is deliberately built only in another environment will be absent here. Tell us, and the finding comes out of your report. Ephemeral models are never judged, because dbt never builds them as a relation.
What we need to check it
manifest.json and catalog.json, generated together by dbt docs generate. A catalog made with --select, covering under 80% of models, is too partial to use, so the rule declines.