MetricSign
EN|NLRequest Access
Low severityoperational

Power BI Refresh Error:
dbt docs generate failure

What does this error mean?

The `dbt docs generate` command failed to produce a complete catalog.json file, either because dbt could not query the warehouse's information schema, encountered a timeout while introspecting large schemas, or ran into a permissions error on system catalog tables. The docs site may be missing column-level metadata or fail to build entirely.

Common causes

  • 1The dbt service account lacks SELECT on information_schema or SHOW TERSE OBJECTS in the target database (Snowflake-specific)
  • 2The target database has a very large number of schemas or tables, causing the catalog introspection query to time out
  • 3A dbt Cloud job step runs `dbt docs generate` but the warehouse connection times out during schema introspection
  • 4The `--select` flag used with `dbt docs generate` resolves to zero nodes, producing an empty catalog
  • 5dbt version incompatibility with the warehouse adapter causes the catalog introspection query to fail silently

How to fix it

  1. 1Run `dbt docs generate --debug` locally to get the full error from the catalog introspection query.
  2. 2Grant SELECT on the warehouse's information schema to the dbt service account (Snowflake: `GRANT IMPORTED PRIVILEGES ON DATABASE <db> TO ROLE <role>`).
  3. 3Use the `--select` flag to limit catalog generation to only the models in the project, avoiding full-schema introspection: `dbt docs generate --select <tag_or_model_name>`.
  4. 4Increase the warehouse size for the docs generation step to avoid query timeouts on large schemas.
  5. 5If docs generation is blocking the job on timeout, move it to a separate non-critical job step that does not fail the main transformation job.

Frequently asked questions

Does a dbt docs generate failure stop my models from running?

No — if `dbt docs generate` is a separate step in the dbt Cloud job, a failure in that step does not prevent `dbt run` or `dbt test` steps from executing, provided the job is configured to continue on docs failure.

How do I speed up dbt docs generate on a large schema?

Use `dbt docs generate --select <tag>` to limit introspection to only the models in your project. On Snowflake, using a larger warehouse size for the docs step also significantly reduces catalog query time.

Other operational errors