... you should ramp up slowly when load testing an application on App Engine. Ramping up too quickly won't give an accurate picture of how App Engine scales; you have to accomodate our load balancing code ...
The problem with this is that the test is written to accommodate known limitations of the architecture. Performance tests need to take the opposite approach. In this case that means specifically revealing the performance implications when App Engine scales up and down.
That said, I don't expect a Google engineer to reveal App Engine's limitations :]
It depends on what your load test is attempting to prove.
Are you attempting to prove you can handle a /.,digg,fark,reddit, etc without a single 500 error in a 10min window?
Or are you attempting to prove that you can handle 2million "transactions" an hour, but have the luxury to ramp up in the real world slowly? (Twitter didn't go from nothing to everything in a day)
I'd say the second case is more likely to be the case 90% of sites would care about.
I don't understand why this is downmodded, a pull-based chat is a pretty good test for load handling capabilities. Seems like this is sending a request every ~2 seconds for every user logged in.
How are you finding GAE so far? BTW are you hitting the daily request # limit?
The problem with this is that the test is written to accommodate known limitations of the architecture. Performance tests need to take the opposite approach. In this case that means specifically revealing the performance implications when App Engine scales up and down.
That said, I don't expect a Google engineer to reveal App Engine's limitations :]