Performance & Benchmarking


#1

We are planning to setup a benchmark with up to 4 Cafienne/Akka nodes.
The purpose is to find out how the engine scales. Also we know of some sharding parts of the engine that are yet to be implemented. It should pop up in this setup, and the setup then also provides for a means to fix.

Basic idea is to have 4 or 5 diffferent case models, each with a “unit” testing path, and then use e.g. SOAP UI test cases to generate load.

We’re interested in getting feedback on this idea. Tooling, case models, test scripts, debug help, anything goes :slight_smile:

There is no concrete timeline as of now, but we hope we can spend the month of June on this.


#2

Cool stuff.

I’d propose a way that we can automatically create such a setup and have test that can be run automatically based on the setup.

For the load/performance testing, we’ve used Gatling in the past and when you like to take such an approach, I can share a sample project with gatling.