There is quite a few examples on how to autoscale workers and dynes on Heroku, but I could not really find anything on autoscaling Resque workers in EC2.
There is resque-director gem but that's about it..
Is there a solution for autoscaling Resque workers (launching new ones and stopping unused ones), that takes resource management into consideration?
On EC2 you'll need to create autoscaling groups using the shell api or one of various SDKs.
There's also a SDK available for Ruby http://aws.amazon.com/de/sdkforruby/
Autoscaling has to be set up once, then scaling should be out of your mind. Rules for autoscaling are based on metrics like cpu usage, network bandwith and others.
You can define scale out triggers to increase and cooldown triggers to reduce the amount of servers.
More details you can find here http://aws.amazon.com/en/autoscaling/
Here's a simple tutorial which describes the steps http://www.caseylabs.com/how-to-setup-auto-scaling-on-amazon-ec2