Address known failure scenarios

[Mierdin]

We have a number of "known" failures that we could do a better job of automating. For instance, the bootstrap script timeout thing keeps happening, maybe just increase the timeout? Also, sometimes I need to manually delete a CNAME in AWS - need to figure out why this is happening and make sure it's automated

Note that this is not just a st2cd issue, we can do this in st2ci as well