May 29, 2015 at 2:41 pm #8277@punkmikOnly available when logged in
I am learning a bit about performance testing recently and thought I would share how we approached a load testing need/problem and to hopefully get some more tips and tricks. 🙂
The performance test we did was for a re-launch of a website.
First step was to define reasons as to why we wanted to load and performance test the website:
1. Ensure new infrastructure is reliable and responsive
2. Find maximum capacity so as to respond adequately to traffic spikes
We then set Load Testing objectives:
1. Handle up to 1000 users per hour (~20% more than the greatest load experienced so far).
2. 200 users should be able to use the site simultaneously.
3. Error rate below 1%
Then we used existing google analytic metrics to analyse a typical user journey through the website and blog, for example:
Blog: The average user reads 1.3 articles, so the test will ensure 2 articles are read.
In this journey the test will:
Navigate to the root domain
Navigate to /blog
Navigate to an article
Navigate to a related article
Leave the site
We had a couple of these scenarios ranging from simple as above to more complex.
Following those tests we added 2 more:
1. A final test journey is a random selection of URLs from the sitemap to more accurately represent user indecisiveness and randomosity.
2. Programmatically; load the sitemap XML, parse it, create HTTP requests
Issues we had:
1. Local machines were bottlenecks – JVM and internet could not keep up with large number of requests
2. Development work stopped by excessive testing – slowed the manual test environment down
3. Maths (because threads, ramping up times etc were new to us)
How do you approach load and performance testing? Any tips for a newbie in the field?June 25, 2015 at 3:26 am #8614@rockyOnly available when logged in
Thanks for sharing.
I’m curious to know more about Performance testing…what is your favorite source to learn performance testing?
You must be logged in to reply to this topic.