"Waiting for a an API call could lower CPU" - well, that depends. If you're 
running in threadsafe mode enabled then yes, maybe.

Total CPU means exactly what it says: total, i.e. API calls + your app code 
CPU. Say, if you have a Total CPU of 10000ms and the blue line is at 
9900ms, that means only 100ms were used for a service API call. The rest is 
somewhere very close to your app. 

At least, that's my understanding. 

Another example of a (very unexpected) error I've seen quite often is this: 
"'ascii' codec can't decode byte 0xe9 in position 0: ordinal not in 
range(128)". This, again, error in the app code. The developer completely 
forgot that we have different encodings. The app was working just fine for 
months, and it just happend that one day a non-US user signed up with a 
name that contained non ascii chars.

Have you tried to run your app with AppStats and see where it's taking most 
of the CPU?


On Wednesday, March 14, 2012 6:50:10 PM UTC+1, Mos wrote:
>
> Thanks for your thoughts....
>
> > In fact, in your graphs the API calls CPU is as low as it always was
>
> CPU for API calls is not related to latency and timeouts, right? To the 
> contrary:  Waiting for a an API call could lower CPU, doesn't it?
> --> It seems not to help for this analysis
>
> >  ... using (e.g. URLfetch). ...  re-populating (re-caching)
>
> No nothing of these.
>
> > Well, I don't personally think that every filed ticket means a real 
> issue ...
>
> Please check the Outage-Thread and the other posting the last hours/days. 
> It's really obvious that something is going on.....
>
> Cheers 
> Mos
>
>
>
> On Wed, Mar 14, 2012 at 6:38 PM, alex <a...@cloudware.it> wrote:
>
>> Well, I don't personally think that every filed ticket means a real 
>> issue. It often happens that the issue is in the code or something else, 
>> unrelated.
>>
>> In fact, in your graphs the API calls CPU is as low as it always was. 
>> That means the cause is neither Datastore, nor Memcache or any other 
>> service your app might be using (e.g. URLfetch). 
>>
>> Most likely that means the actual code of your app is consuming the CPU, 
>> so the scheduler simply launches more instances because the average latency 
>> of your app has increased. 
>>
>> Even though the traffic to your app is the same as it was before, it very 
>> much might be that your code is not covering all edge cases. I don't know, 
>> some missing entity property that happend to be there only in the past 24 
>> hours, or maybe you rely on memcached query results that are not there 
>> anymore and you just not re-populating (re-caching) some of it. Stuff like 
>> that.
>>
>>
>> On Wednesday, March 14, 2012 6:15:47 PM UTC+1, Mos wrote:
>>>
>>> Hi Alex,
>>>
>>> right, 1 instance is billed, cause this application runs in non payed 
>>> mode, but the scheduler spans up to four instances (without real need).
>>> In payed mode I assume the green line would not be on 1 instance all the 
>>> time...  (I'm glad I didn't activate payment yet)
>>>
>>> Graphs are attached.
>>>
>>> According to Google Analytic and application logging there was similar 
>>> usages then the days before.  (Around one to two request the minute)
>>> I think we don't need to discuss that something bad is going on the last 
>>> 24h to 48h hours for many GAE users?
>>> Please check issue tickets and other postings here.....
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Wed, Mar 14, 2012 at 5:58 PM, alex <a...@cloudware.it> wrote:
>>>
>>>> Funny... maybe I'm not getting something here but, where's the issue? 
>>>> It is crystal clear from your graphs that you are being billed for 1 
>>>> instance only, all the times.
>>>>
>>>> Can you show CPU ms used/sec and Errors/sec graphs?
>>>>
>>>>
>>>> On Wednesday, March 14, 2012 5:50:04 PM UTC+1, Mos wrote:
>>>>>
>>>>> And here are two that are related to this thread  (application has 
>>>>> same traffic, load and configuration for days  --> major GAE problem for 
>>>>> more then 24 hours):
>>>>>
>>>>> Attached....
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>> On Wed, Mar 14, 2012 at 5:29 PM, Brandon Wirtz 
>>>>>> <drak...@digerat.com>wrote:
>>>>>>
>>>>>>> Nah it is all shiny… And that couple of hours when there are no 
>>>>>>> logs, that was just cause everyone was sleeping… 
>>>>>>>
>>>>>>> This is just a pretty picture… It’s not related to this thread. J
>>>>>>>
>>>>>>>  
>>>>>>>
>>>>>>>  -- 
>>>>>>> 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 <google-appengine@googlegroups.com>.
>>>>>>> To unsubscribe from this group, send email to 
>>>>>>> google-appengine+unsubscribe@**g**ooglegroups.com<google-appengine%2bunsubscr...@googlegroups.com>
>>>>>>> .
>>>>>>> For more options, visit this group at http://groups.google.com/**
>>>>>>> group**/google-appengine?hl=en<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 view this discussion on the web visit https://groups.google.com/d/**
>>>> msg/google-appengine/-/tk8Xo_**fvQ04J<https://groups.google.com/d/msg/google-appengine/-/tk8Xo_fvQ04J>
>>>> .
>>>>
>>>> To post to this group, send email to google-appengine@googlegroups.**
>>>> com <google-appengine@googlegroups.com>.
>>>> To unsubscribe from this group, send email to 
>>>> google-appengine+unsubscribe@**googlegroups.com<google-appengine%2bunsubscr...@googlegroups.com>
>>>> .
>>>> For more options, visit this group at http://groups.google.com/**
>>>> group/google-appengine?hl=en<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 view this discussion on the web visit 
>> https://groups.google.com/d/msg/google-appengine/-/9EjTUlYsfnQJ.
>>
>> 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.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/google-appengine/-/4x6popTT7REJ.
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