What are best practices for dev-staging-prod envir...
# best-practices
s
What are best practices for dev-staging-prod environments using dbt and meltano (I'm using bigquery, but this is a general question)? Do you use a project for raw data and a project for results? In which project do you test new taps?
v
Gave me the 🤯 I needed with DBT , hope it helps you!
s
Thanks!
v
But maybe that's not the question you have here so I'm just guessing
s
I'll definitely look into it; I need all the dbt training I can get. My question is more aimed at how you setup your production and dev environments in your target, and how do you implement new meltano taps / dbt models
c
My first guess is that the practices for dev-test-prod and related topics will vary a bit depending on the scale and velocity of your specific environment (and the data in question)
s
Yes this is what I'm realizing; I was working blind trying to organize my data, but I think I'm going to follow the dbt suggested approach. Update @visch indeed, this tutorial is blowing my mind 🤯
c
My preferred approach: Iterate quickly! i.e. Do what works. Learn as you go. Fix problems as they arise. e.g. Regarding dbt project structuring, I was aware of the excellent guidelines on dbt from the blogs, but started without reading all of them to show value of dbt. As my dbt projects grew, I quickly started adopting a lot of the recommendations from the blogs that Derek linked.
s
That's probably the best approach, especially with the ever evolving landscape 😉 thanks @christoph!
p
@Stéphane Burwash if its helpful for you, I actually created a

walkthrough video

of our isolated development environments last week! It allows us to have environment and developer level isolation. We use snowflake but most of the configuration is just toggling schema/database names based on meltano environment configs so it should be pretty generically useful
s
Awesome thank you so much! I'll add it to my to-watch list 😉 I wanted to commend you on your videos btw; I very much enjoyed the dbt-jaffle shop one
p
Thanks! I'm glad theyre helpful