esme-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ethan Jewett <>
Subject Re: Anybody doing load tests on the test stax instance yesterday?
Date Tue, 15 Dec 2009 13:37:31 GMT
Also, my Growl client prototype was making requests every 2 seconds
for several hours on Sunday because I forgot to change the long-poll
timeout back to 300 seconds. Also, if requests from it are not
successful it doesn't currently behave very well and just keeps trying
to request messages as fast as its little legs can run. Need to fix


On Tue, Dec 15, 2009 at 2:58 AM, Markus Kohler <> wrote:
> Hi Dick,
> I'm guilty of course ;-)
> I played around with a test which would send messages with several users in
> parallel. It's not fully working yet, but I got some hints in the mean time,
> why it is not.
> I "forgot" to add a log off step, ok I was lazy. That probably explains the
> high number of sessions. The "response time" is very hard to measure for
> ESME because the POST that sends the message is usually very quick, but when
> the message shows up somewhere is hard to detect.
> Markus
> "The best way to predict the future is to invent it" -- Alan Kay
> On Tue, Dec 15, 2009 at 5:29 AM, Richard Hirsch <>wrote:
>> Was anybody doing load tests yesterday on the
>> instance?
>> I just have a few bits of information since the display of Stax logs
>> is restricted to a certain length.
>> Sessions went from 0 to 40,000 in less than an hour.  We also had
>> 1,800 requests in the same time period. Memory was high at 333 MB but
>> CPU utilization was low at 15% during the tests and rose later to 25%.
>>  Request time remained low at 20 seconds but this doesn't tell me much
>> since I don't know what requests were being used.
>> D.

View raw message