Hello Team,
I am working on POC about k6 tool and cloud for my organization. I would like to know, over k6 cloud, when I select any region, which mechanism is used to allocate load generator server within that region? If all load generators are occupied within particular region, then what happens to my test?
Thanks in advanced!
I am sorry, I don’t understand what exactly are you asking here. If you choose only a single region for all of your load generators, then that’s where all of the instances will be spun up and where all of traffic for your test will be coming from.
You can configure that from both the test builder and via the CLI / script editor options, see:
Hello @ned ,
Thank you for your reply. I understand what you are saying. But, let me reframe my question.
1.Every-time, any user trigger the test on cloud platform, will it creates new instances of load generator on that region?
2.If multiple testers trigger test on same region at same time, for some of theme, do their tests queued up waiting for instance to get free?
3.And internally, does it work on master-slave mode? or how the load is distributed among multiple load generators in same region?
Regards!
Regarding points 1. and 2., tests can use newly allocated instances or, for better startup speeds, they can reuse previously allocated instances that are currently free. But a single load generator instance is not shared, it can’t be used by more than one user or load test at any given time, i.e. other tests won’t interfere with your tests.
Regarding 3., it’s a bit complicated and confidential exactly how we currently distribute load and orchestrate tests in our commercial cloud service, so I can’t get into the details. I can say it’s not exactly a tradition master-slave approach, we instead heavily depend on execution segments to eliminate the need for runtime synchronization among instances.