On Monday, June 4, 2012, vivek wrote:
Hi,
> To load test I used loadimpact.com and the results of which can be found on:http://loadimpact.com/load-test/www.reviews42.com-18774e46e8f562a6eb4...
> The test configuration consisted of 600 VUs with 10 mins step duration.
> Got around .1 millions requests and around 200+ requests/sec max. Is this
> good, bad, or at par?
Some quick observations from results:
1. The user load times start from 30+ seconds
That's aggregated load time, and not a single page loading time. The test comprised of navigating to multiple pages to generate more real life scenario.
2. The error rate also increases with higher requests/sec.
Yes. With high reqs/sec it's starts to come in. An upcoming major client side code optimization is in pipeline. A drastic change in reqs/sec is expected.
3. text/html , which is the output of django app, is taking 62.74 %
time.
This number might not be bad actually, taking into consideration that I aim to reduce the number of http connections per page to something pretty low.
Overall things don't seem optimal. If you are testing based on load
times of 30+ sec then the test approach may not be practical, you dont
expect your visitor to wait 30+ sec. Your application itself may need
more fine tuning. text/html generation time seem on higher side.
Also be mindful that generally database performance on Amazon EC2 is
lower compared to bare metal servers but you have mentioned that the
queries are minimal so not sure how much would that have any impact.
As I said above the 30+ secs is the aggregated number. That too for un-optimized browser code.
What is the payload of your html page ?
5- 10 Kb (compressed) on avg depending upon page content
Without knowing much about the application its usually difficult to
say much but based on the results, there seems to be scope of
improvement in html generation itself.
Once the major client side optimization comes through, and some deployment/setup changes, I expect the request/sec handling capability to easily shoot the sky.
Since you thought the aggregated load time to be of a single page, I guess your perspectives need to change accordingly. :)
--
rgds
vivek
--
You received this message because you are subscribed to the Google Groups "Django users" group.
To post to this group, send email to django-users@googlegroups.com.
To unsubscribe from this group, send email to django-users+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/django-users?hl=en.
You received this message because you are subscribed to the Google Groups "Django users" group.
To post to this group, send email to django-users@googlegroups.com.
To unsubscribe from this group, send email to django-users+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/django-users?hl=en.
No comments:
Post a Comment