Types of metadata

Events

Events store timeline data, such as jobs, dashboard views and dbt invocations, model runs and tests. The table loosely follows the idea of One Big Table. There are a lot of fields here, some are generic, such as the time the event happened, and the user associated with the event. There are also platform-specific fields, that will be prefixed as follows:

Looker	lk_
BigQuery	bk_
Snowflake	snowflake_
dbt	dbt_

The easiest way to get started with the events table is to start with example queries and modify them or enter the ↗ Query Editor button from anywhere in the product. This will allow you to understand quickly how to interact with the model to craft your custom insights.

Lineage

The linage table stores the lineage as edges containing from -to as well as additional metadata from the context such as job ID's, query types and timestamps.

Entities

Entities store all entities from all the connected platforms: tables, columns, schemas, dashboards and dashboard elements. This table is not recommended to use - as the entity_search table will fullfil most of the use cases in a more performant way. It's still available if for some reason it could be useful.

Entity_summary

Entities store aggregated information for the last 30 days about an entity, such as number of rows, number of queries, cost and others.

Entity_search

This table is a specialised table that allows fast lookups on name, description, tags. As the name implies it's also used to power the Search feature in Alvin.