Comment 6 for bug 1244766

So I just tested bootstrapping with trunk @ rev 2034 which is just before my Uniter address cache patch. If you tail the /var/log/juju/machine-0.log file you can see lines like:
  2013-11-07 07:30:43 DEBUG juju state.go:160 waiting for DNS name(s) of state server instances [i-1ffaea7b]

Which are triggered 2x every 5s (when you are running 1-x1 and 1-x2).

After running "juju upgrade-juju" with trunk @ 2036 (my patch was 2035) they are no longer in the lag.

I then turned up the number of units (I just deployed all units to the same machine so I wouldn't have to worry about extra cost). And things kept up cheerily. (And I could 'juju status' which *does* currently use the ec2 API to find out instance status.)

Certainly the log file ends up with lots of churn as these hooks are firing over and over again, and it does put a fair amount of load on an m1.small (load average 2.5). But I was happily running away with six x1 and six x2 units

I was able to get up to 16 units of each, though I think at that point the limiting factor was that there is only 1 hook running per machine at any given time.