# aws


03/12/2021, 1:06 AM
Advice for best strategy: Our team wants to use AWS Airflow (which is quite new). Pulumi doesn’t yet accomodate for that service, so we have decided to do the following: 1] Build another separate repository and corresponding Pulumi stack that handles only the provisioning and management of AWS Airflow through CloudFormation. 2] Keep our (mostly) monolithic project in its own repository and Pulumi stack BUT still be able reference and use some of the AWS resources (i.e. S3, VPC) made from #1 after all resources have been entirely and successfully deployed (via
pulumi up
). So, ideally #1 would occur first with its own
pulumi up
and then - after that entire deployment is complete - automatically do the same for #2 but somehow be able to be aware of and use resources that were changed/deleted/added from #1. Finally, being able to “tear” all this down in an automated way would be fantastic, but that’s secondary for now. What your advice for approaching this use-case?