I'm sure either, I inspected the keys in etcd, and they don't correspond to events.
I think we'll start over with this new cluster, we have rollbacked to 1.0.6 for now, we have some deploys to do today.
I will sent you some details in the coming days if we have the same issue.
Currently, we can't do anything, the master is restarting every minutes, and even the nodes are behaving weirdly. node-1 is panic after 2s, while the other nodes don't do this, with the same configuration...