It could just be some rogue deployment script running from EC2 that are a little more active that it should be. Imagine someone is deploying their 1GB repo from GitHub to 100 small EC2 instances :)
Initially we blocked all Ec2[1] & spamhaus ip list. But then realized Flipboard proxies[2], some blog aggregation proxies etc are based on Ec2 machines.
What would be a good way to block such rogue machines? Is there a community sponsored list or Ec2/Rackspace ips that are creating issues?