Hi Greg,

On Thu, Aug 12, 2010 at 10:14 PM, Greg Tracy <[email protected]> wrote:

> Here's an example of what I mean by those metrics being inversely
> proportional. Each of these tasks are doing the identical work and ran
> two minutes apart.
>
> 08-12 01:42PM 11.860 /crawl/prefetch/03 200 6227ms 108803cpu_ms
> 107111api_cpu_ms
> 08-12 01:39PM 47.970 /crawl/prefetch/03 200 23846ms 25655cpu_ms 23770
> api_cpu_ms
>
> The following are my measurements using the quota package and they
> represent the end-to-end quota times inside my handler for these
> respective calls...
>
> 08-12 01:42PM 18.056 total CPU cycles 5584, total API cycles 128534
> 08-12 01:40PM 11.517 total CPU cycles 4132, total API cycles 28524
>
> From an analysis perspective, I have two challenges:
>
> 1. I cannot reconcile the wall clock time for these slow requests
>

How are you measuring the wall clock times?


> 2. I cannot trust the metrics from the quota package since I can't see
> clarity about what they actually represent
>

You're comparing milliseconds on the one hand to CPU megacycles on the
other. The quota package contains functions to convert between the two.

-Nick Johnson


>
>
>
> On Aug 12, 12:55 pm, Greg Tracy <[email protected]> wrote:
> > In fact, there may even be a bug... I see that the cpu_ms and
> > api_cpu_ms numbers are inversely proportional to the wall clock times
> > in the log.
> >
> > I can't make heads or tails from the existing reporting tools. Please
> > save me from this insanity... :)
> >
> > On Aug 12, 12:27 pm, Greg Tracy <[email protected]> wrote:
> >
> >
> >
> > > Excellent... Now how do I reconcile the times reported by the quota
> > > package with the "cpu_ms" and "api_cpu_ms" times reported in the log?
> >
> > > They don't appear to be the same.
> >
> > > On Aug 12, 10:50 am, "Nick Johnson (Google)" <[email protected]>
> > > wrote:
> >
> > > > Hi Greg,
> >
> > > > On Thu, Aug 12, 2010 at 4:36 PM, Greg Tracy <[email protected]>
> wrote:
> >
> > > > > Is the billable "CPU Time" quota a combination of the "cpu_ms"
> cycles
> > > > > and the "api_cpu_ms" cycles (as seen in the log file)?
> >
> > > > Yes.
> >
> > > > > And if this is the case, can I assume that "cpu_ms" is measuring
> the
> > > > > time my code is executing - unblocked from remote calls?
> >
> > > > Yes.
> >
> > > > -Nick Johnson
> >
> > > > > --
> > > > > 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]<google-appengine%[email protected]><google-appengine%2Bunsubscrib
> [email protected]>
> > > > > .
> > > > > For more options, visit this group at
> > > > >http://groups.google.com/group/google-appengine?hl=en.
> >
> > > > --
> > > > Nick Johnson, Developer Programs Engineer, App Engine Google Ireland
> Ltd. ::
> > > > Registered in Dublin, Ireland, Registration Number: 368047
> > > > Google Ireland Ltd. :: Registered in Dublin, Ireland, Registration
> Number:
> > > > 368047
>
> --
> 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]<google-appengine%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/google-appengine?hl=en.
>
>


-- 
Nick Johnson, Developer Programs Engineer, App Engine Google Ireland Ltd. ::
Registered in Dublin, Ireland, Registration Number: 368047
Google Ireland Ltd. :: Registered in Dublin, Ireland, Registration Number:
368047

-- 
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