Did anyone else run into issues with ELB during the outage? We're multi-AZ and could access unaffected instances directly without a problem, but the load balancer kept claiming they were unhealthy.