On Fri, May 13, 2011 at 3:29 AM, Brandon Wirtz <drak...@digerat.com> wrote:
>
> I am slow to upgrade because I like to know that things won’t explode, but 
> this graph
> shows how 1.5 reduced the number of instances I need compared to 1.4
> (maybe 1.3) that I was running.  Usually I version thing so you can’t see the 
> graph before.


I don't think that's how versioning works.  Google updates their
infrastructure independently of which SDK you have installed on your
laptop. You can see this effect as people report fresh bugs in their
deployed apps without having redeployed.

So delaying updating your SDK is not the conservative strategy. To be
conservative, you should install the preview SDK as soon as it is
released and run your tests locally. The production machines seem to
be updated with the preview SDK release, which is about 2-7 days
before the final SDK is announced.

So why the change in number of instances? I don't think you can read
too much into it. Under the current pricing scheme Google's incentive
is to optimise the scheduler because they pay for any inefficiencies.
They can't run every machine at 100% utilisation all the time. It may
be the case that the machines your app happened to be deployed to last
week had spare capacity and Google calculated that it was more
efficient to leave more of your instances relatively idle in memory
that to stop and restart them. This week, perhaps some collocated apps
had a traffic spike and Google calculated that it would be more
efficient to allocate some of those instance slots to someone else.

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to google-appengine@googlegroups.com.
To unsubscribe from this group, send email to 
google-appengine+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en.

Reply via email to