Say I have 200 ec2 instances I want to roll from A to B. Is there a way to limit the number in-flight ec2 instances that Pulumi will build/destroy at once? (Rate limiting the pulumi API isn't an option due to the number of calls needed to support the app.) Ideally what we want is for Pulumi to build 1, destroy 1, build 1, destroy 1 ... all night long.

are they in an autoscaling group? or have you defined 200x instances in a loop?

200x instances are in a loop

pulumi up --parallel 1

will do it, but please test it first

thanks - good suggestions