Comment 4 for bug 1017418

Ah, Thierry - I forgot to add this little ditty which is crucial to this. I modified the rate limit - as well as quotas (for internal clouds, the quotas are too strict when the idea is, under an account, to spin up as many machines as required - especially for a "production" tenant)...:

/etc/nova/api-paste.ini:
[filter:ratelimit]
paste.filter_factory = nova.api.openstack.compute.limits:RateLimitingMiddleware.factory
limits =("POST", "*", ".*", 120, MINUTE);("POST", "*/servers", "^/servers", 1000, DAY);("PUT", "*", ".*", 120, MINUTE);("GET", "*changes-since*", ".*changes-since.*", 30, MINUTE);("DELETE", "*", ".*", 100,MINUTE)

In terms of DoS and billed: for internal private clouds, this very probably won't involved money changing hands so this makes the resource "limitless" providing the local network level DoS

I don't think ratelimit or quota will help this race condition.
I've not seen floating ips re-appear after a while - after returning the next day to the environment, the environment is still broke.