04. Advanced Environment Strategies
️ Advanced Environment Strategies in dbt
Managing multiple environments effectively is crucial for robust CI/CD, testing, and production deployments.
Advanced environment strategies help ensure isolation, consistency, and reliable deployments across dev, staging, and production.
Key Environment Strategies
Environment separation – dev, staging, prod schemas or databases
Dynamic profiles – configure targets dynamically based on environment
Feature branches – test new models without impacting production
Environment-specific variables – control behavior per environment
️ 1. Using Profiles for Multiple Environments
Define multiple targets in profiles.yml:
my_project:
target: dev
outputs:
dev:
type: snowflake
account: <account>
user: <dev_user>
password: <dev_password>
role: <dev_role>
database: my_db
warehouse: my_wh
schema: dev_schema
prod:
type: snowflake
account: <account>
user: <prod_user>
password: <prod_password>
role: <prod_role>
database: my_db
warehouse: my_wh
schema: prod_schema
- Switch environments via
--targetin dbt CLI:
dbt run --target prod
️ 2. Environment-Specific Variables
- Define variables in
dbt_project.yml:
vars:
dev:
data_refresh_days: 7
prod:
data_refresh_days: 1
- Access in SQL using Jinja:
select *
from {{ source('app', 'orders') }}
where order_date >= dateadd(day, -{{ var('data_refresh_days') }}, current_date)
️ 3. Feature Branches and CI/CD
- Use Git feature branches for development and testing
- Merge into main or release branches after CI/CD validation
- Ensures production stability while allowing iterative development
️ 4. Best Practices
Keep environments isolated
Use environment-specific variables for flexibility
Automate deployment via CI/CD pipelines
Monitor environment performance and resource usage
Test incremental models and snapshots in staging before prod