dbt is a tool used by several data teams to define and implement their core data models. However, dbt has lacked a mechanism for users to specify metrics- time series aggregations over tabular data supporting dimensions. Here, Drew Banin (dbt Labs co-founder) defines the specification and functional requirements for metrics, which will be implemented as a distinct node type in the dbt DAG. In the future, dbt Core will expose these metric definitions (which can be documented and version controlled) to downstream tools for evaluation and analysis.