Databricks Billing

This package was built to help you monitor the costs involved in using Databricks as a data plataform. It makes it easier to keep up with the costs related to DBUs and clusters.

Before creating a branch

Pay attention, it is very important to know if your modification to this repository is a release (breaking changes), a feature (functionalities) or a patch(to fix bugs). With that information, create your branch name like this:

• release/<branch-name> or major/<branch-name> or Release/<branch-name> or Major/<branch-name>
• feature/<branch-name> or minor/<branch-name> with capitalised letters work as well
• patch/<branch-name> or fix/<branch-name> or hotfix/<branch-name> with capitalised letters work as well

Revisions

0.3.0 - For Snowflake warehouses 0.3.1 - For Databricks warehouses

Quickstart

Requirements

dbt dbt version >= 1.0.0

dbt_utils package: dbt-labs/dbt_utils version: 1.1.1

Install the package

Include the following package version in your packages.yml file.

packages:
  - git: https://github.com/techindicium/dbt-databricks-billing
    revision: # 0.3.0 or 0.3.1

Then run dbt deps to finish the setup.

About the sources

This package uses four main sources:

• list_prices
• usage
• warehouses
• clusters

list_prices and usage are system tables of Databricks, located inside system.billing.

The warehouses and clusters tables contain informations cuptured by a REST API request.

More information about the endpoints could be finded here:

Databricks REST API endpoints

Warehouses endpoint

Clusters endpoint

You can use our adf tap to extract this informations:

Platform Meltano on Databricks

Define database and schema

The location of the raw data to be used in this package is configurable, so it's importante to complete the following information at dbt_project.yml:

models:
  databricks_billing:
    marts:
      +materialized: table
    staging:
      +materialized: view

vars:
  databricks_billing_database: # name of the database
  databricks_billing_schema: # name of the schema

In this case, it's important to put the name of the catalog and schema where the tables warehouses and clusters are.

Recommendation

We strongly recommend that the job that will run this package to be apart from the job that runs the models in production.

This allows to prevent trubles with errors with the package that could make models in production to crash.