Member-only story
How to Deploy dbt to Production using GitHub Actions
How to (and not to) deploy dbt to production.
With the rise of the Modern Data Stack, more and more people use dbt as the main tool for data transformations, aka data modeling. The folks at Fishtown create an amazing dbt Cloud offering that suits the needs of data teams, large and small. With dbt Cloud, any Analyst, seasoned or fresh, can easily start modeling and deploying data transformations pipelines to production.
I highly recommend checking out dbt Cloud since the product does much more than helping you deploy dbt. Firstly, let’s look at how you should not deploy dbt to production.
What not to do
There are many ways to deploy dbt to production. Not all of them are good ways, though. Here are some not so good ways to do that:
Spinning up a compute instance?
A common approach is to spin up a compute instance and install the required packages. From here, people can run a cron job to do a git pull and dbt run on a schedule.
