This message was deleted.

they are basically entirely separate jobs

Everyone does this stuff differently, and being able to use a programming language certainly makes it easier to apply this kind of environment/stack-specific logic. But the best practice is probably to use configuration to capture the differences between environments, and use a single program to cover all of them. (So not much logical difference between them.)

for now i've just taken the tack that if code belongs in one environment, but not another, it's a different project. so here i just created a new "project" for the pipeline itself