Btw, I have a suggestion for you in optimizing for the 5 minute hourly spike
you experience.  If you want to always ever have only a single instance, and
can tolerate high latencies during that 5 minute window every hour, what you
can do is use pull mode task queues.  In that 5 minute window, all your
requests can be queued up in a pull queue (call it the spike queue).  Then
your single instance leases these requests one at a time, and processes them
one at a time.  If your "MaxInstances" parameter is set to 1, and this
single instance is busy serving the spike at the start of the hour, your
other requests might get delayed.  To fix that you can put all other
requests in a different pull queue - call it the high priority pull queue.
 Your single instance should now check for entries in the high priority pull
queue after processing every entry in the spike queue.  So your regular
requests will at most incur a latency of however long it takes to process a
single request in the spike queue.

Hope this helps.

Rishi

On Fri, Sep 9, 2011 at 1:13 AM, Tammo Freese <[email protected]> wrote:

> Missing in the last post:
>
> (2) http://groups.google.com/group/google-appengine/msg/1a2d27d15781ddb7
>
> --
> You received this message because you are subscribed to the Google Groups
> "Google App Engine" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected].
> For more options, visit this group at
> http://groups.google.com/group/google-appengine?hl=en.
>
>

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

Reply via email to