Re: appstats analysis - Same request - Too much difference in serving time.

[Rajesh Gupta]

Hi,

I have posted this post, with attachments twice.  But it did not make into the group.  Any limitation on posting with attachments.  Any settings to be enabled.

I have problem with serving the same request with big difference in the datastore and memcache timings.  Please see the below email.

On Sat, Aug 1, 2015 at 10:12 AM, Rajesh Gupta <rajesh...@veersoftsolutions.com> wrote:

resending... as it did not post earlier.

Attached are the two screen shots showing the appstats

- same request

- same appengine instance

- same user

The first request took 2914 ms, and the second request takes only 1032ms

As you can see in the attached images, 

3rd line datastore_v3.RunQuery, difference of 9ms 

5th line memcache.get, difference of 73ms

6th line memcache.get, differnce 73ms

Through out the request, there is lot of time difference in the datastore and memcache calls, between the two requests.

What can it done to make consistent.

--

Rajesh

www.VeersoftSolutions.com

www.AccountingGuru.net

Accounting/Inventory/Orders on Google Cloud Platform and Mobile

--

Rajesh

www.VeersoftSolutions.com

www.AccountingGuru.net

Accounting/Inventory/Orders on Google Cloud Platform and Mobile

--

Rajesh

www.VeersoftSolutions.com

www.AccountingGuru.net

Accounting/Inventory/Orders on Google Cloud Platform and Mobile

[Rajesh Gupta]

[Rajesh Gupta]

resending... as it did not post earlier.

[Alex Martelli]

On Mon, Aug 3, 2015 at 4:43 AM, Rajesh Gupta <rajesh...@veersoftsolutions.com> wrote:

Hi,

I have posted this post, with attachments twice.  But it did not make into the group.  Any limitation on posting with attachments.

Sorry Rajesh, it was a temporary glitch with the group -- I have now unblocked your messages, and future ones from you should just go through (sorry I don't know enough to actually help w/your question, but we can hope somebody else on the group can).

Alex

 

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-appengi...@googlegroups.com.
To post to this group, send email to google-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/google-appengine.

To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/CA%2BS7ijYGLJvcPK8Cfi7Vmv%2B%3Dekx5BqyVxPeMdKt-o9K9Hdc1vw%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

[Jason Collins]

Rajesh - if you're performing exactly the same execution and are simply asking about the variability - that's just sort of the nature of the distributed RPCs. There can sometimes be pretty wide range of response times in RPCs.

Depending on your frameworks, there are ways to execute queries in parallel which may help. Also, your framework may be trying to "help" by automatically using memcache (e.g., NDB does this, though not for query results); if you memcache hit rate is low, then all the memcache overhead is just waste.