Hi -- exploring Pulumi. I really like the design....
# general
Hi -- exploring Pulumi. I really like the design. I'm trying to figure out if I can use it to support my development workflow for a personal project. I need to spin up gpu instances in the cloud for running code in anaconda environments with gpu support. I've been manually setting up machines and rsync'ing data (its feasible to rsync data for my problem size) to them to support development. I'm wanting to see if I can build an abstraction on top of (or composing) aws.ec2.Instance to manage a portion of the filesystem on the instance(s) via (something like) rsync. Is the right thing to just use pulumi to manage the lifecycle of the instance(s) and then interrogate pulumi with
pulumi stack output --json
from another script to figure out how to contact the hosts and drive the rsyncing there? Doers that sound like a possible approach or other recommendations on how to do this? If I'm using pulumi typescript can I get the equivalent of
pulumi stack output --json
somehow without manually invoking the pulumi command and parsing the json output myself -- ideally in a way that preserves type info and auto-complete behavior