Dashboard -> Current Load: please return back Avg Latency & CPU per URL

[Sharp-Developer.Net]

With new pricing we've got a new dashboard and it did not got better.

Until today in the "Current Load" tab we were able to see most demanding handlers that showed average latency & CPU per URL. Combined it was a powerful tool for spotting wrong things. It was so good in identifying performance/money troubles that we even changed some URLs from /resource/ABC to /resource?id=ABC so we can get the accumulated stats at the dashboard. We can not anymore - it shows just  Req/Sec and Total number of request.

Any chance to get Avg. Latency & CPU back?

Anyone feel the same as we do? Should we raise an issue in official tracker?

Alex

[Gregory D'alesandre]

Hi Alex,

Please feel free to raise an issue in the issue tracker.  We are looking into improving the dashboard to allow you to better optimize your app.  I'm interested to understand more about what you would do with the information.  For instance, why is CPU still important for you?  Thanks for the suggestion!

Greg

--
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/-/reL-M3oBW7IJ.
To post to this group, send email to google-a...@googlegroups.com.
To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.

[Kenneth]

Hi Greg,

On a similar vein, I've put in a request for the datastore costs to be put in to the logs:

http://code.google.com/p/googleappengine/issues/detail?id=6258 

Apologies if this is a dup but I couldn't find anything similar.

Thanks,

Kenneth

[Sharp-Developer.Net]

Greg, thanks for prompt reply.

I've raised an issue: http://code.google.com/p/googleappengine/issues/detail?id=6259

Anyone feeling it reasonable please start it!

To answer your question: Why we need both Latency & CPU with the new pricing model?

High latency identify there could be sequence of sync calls to API or improper limit.

High CPU with low latency can identify here can be a case when we have many async request to API (while we should/can use in-instance cashed data).

Does my logic sounds reasonable? At least it works for us. We use this approach to incrementally optimize top handlers one by one. Usually we get 10x improvements - completed 3 iterations at the moment.

[Gregory D'alesandre]

Thanks, that makes total sense, I just wanted to understand how you were using it to ensure we provide you the right information.

Thanks again!

Greg

--
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/-/3SnAwvZyQTkJ.

[Vivek Puri]

+1

On Nov 7, 5:23 am, "Sharp-Developer.Net"

<alexander.trakhime...@gmail.com> wrote:
> Greg, thanks for prompt reply.
>

> I've raised an issue: *http://code.google.com/p/googleappengine/issues/detail?id=6259*

>
> Anyone feeling it reasonable please start it!
>

> To answer your question: *Why we need both Latency & CPU with the new
> pricing model?*

[Strom]

CPU usage isn't that important, but it's better than nothing.
Would be very nice to see the instance time used by every URI.
I used this table to find big resource hogs.

[pdknsk]

If this costs resources, I prefer to have Google just add latency but
not CPU. The dashboard is often quite slow already as it is.