hi all, We're starting to use jmeter on amazon's ec2 cloud to generate load.
Can anyone post tips/best practice/words-of-caution/"this-worked-for-me"? This post here uses 'scalr.net' which provides cloud services on top of amazon ec2. http://vedovini.net/2009/08/jmeter-distributed-testing-with-amazon-ec2/ [I'm not exactly sure what scalr.net is, but it's like "pulling teeth" to get an additional expense approved. So we'd like to stick with ec20 Issues: -is there an easy way to get IP addresses of jmeter slaves (they change upon startup, unless you use elastic ip (i.e. static ip) -cost/benefit of using elastic ip's for testing...to minimize ip address changes -any variance in traffic -cost/benefit of using an extra large amazon instance to generate load vs. several small instances -any easy way to automate the slaves so that upon instance startup they communicate with the master (i..e ala zombie pc's) -virtual private cloud I'm not looking for definitive answers..any insights will do. Please email me privately if you have some notions, but don't have the time to for a fancy writeup for posterity. I will summarize later and post to this thread. Another link: http://www.tikalk.com/flex/stress-test-your-flex-application-using-jmeter-and-ec2-part-1 thanks in advance will
