Skyrocketed instances problem on python apps.

[Christopher Wong]

Hello,

It caught my attention that instance usage on our app (tetris4fun) has been shooting the roof - at some point there were 250+ instances running, and 20 is what we need on daily basis so far.

Attached are the graphs for the instances and the empty instances in question. I have also attached the request graph to show that our usage pattern is consistent.

Thanks,

Chris

[Kyle Jensen]

I am seeing the same thing...HUGE increases in instance numbers for brief periods: up to 500 in my case.

However, it doesn't appear that is reflected in the billing, as far as I can tell.  Something has changed....there are a few posts about this in the forum, but almost no responses.

Kyle

[Christopher Wong]

The spike is back again around 5 hours ago.

On Friday, March 9, 2012 5:03:45 PM UTC-5, Christopher Wong wrote:

[Chris Dimitropoulos]

Same problem here, for most of the days on last week, and especially
on March 13th, from an average of 15 instaces per day, we jumped to 60
instances, as a result of increased milliseconds per request (no
change from our side). I looked up system status and there is a huge
latency increase for March 13:
http://code.google.com/status/appengine/detail/serving/2012/03/13#ae-trust-detail-static-get-small-nogzip-latency

In our case, this is reflected in billing, from an average cost of $10
- $15 per day, we jumped up to $54 per day.

Can pls someone explain what happened and if we deserve a refund or
something? Our budget has really gone through the roof...

tnx

>  instances.JPG
> 170KViewDownload
>
>  instances2.JPG
> 28KViewDownload
>
>  cache.JPG
> 31KViewDownload