I'm looking for a way to detect the active environment using environment variables and I know you can set a meltano environment using

--environment=prod

or through a

MELTANO_ENVIRONMENT

. Its easy to solve if I use the second method but if I define it using the

--environment

argument there doesnt seem to be any environment variable set. Is there a way to consistently do this?

Could make a wrapper that calls meltano, and forces environment to be defined eek . Gross but it'll work everytime!

I think we (I) should add the active environment setting to

src/meltano/core/bundle/settings.yml

so the env var always gets injected into the execution environment. @pat_nadolny I'm curious how you intend to use the detected environment.

@edgar_ramirez_mondragon I'm making a custom DAG generator in the squared repo and I want it to check the active environment to decide if it should generate DAGs using the KubernetesPodOperator or the BashOperator. This would let me turn on and run my pipelines locally and also make sure they generate properly, then that dag folder is also used as input for building our airflow docker image thats sent to EKS. I could have it defined as separate modules (dev/prod) but then im worried about skew between them.

@aaronsteers and I had a conversation yesterday about using the current environment to switch between profiles in

dbt

(rather that creating a new Meltano-specific

DBT_PROFILE

setting and env var) 🤔 +1 for injecting

MELTANO_ENVIRONMENT

into plugin context as I think it is much neater and more reusable than a per-plugin setting definition 🙂

@pat_nadolny - Can you create an issue on this? I'm inclined to inject

MELTANO_ENVIRONMENT

into child processes if environment is declared another way (e.g. via the CLI flag, or via default_environment in meltano.yml)

@aaronsteers yep I created one yesterday https://gitlab.com/meltano/meltano/-/issues/3377

I've added comments and added a one-liner spec proposal to the top of the description. Thanks for opening!