We just had 100s of EC2 instances generate high (alleged) load. Instances had load averages of 90+ but were responsive.

Running on a 3.2 kernel

Rebooted them all and they're fine.