Stresscloud is a distributed resource stress manager.
It provides execution of a distributed test scenario to registered stressers on the network.
Once a scenario is performed, you can retrieve performance values
The scenario are written in groovy
- to provide smaller scenarios thanks to collections operators
- to have an easy interpreter
An example of scenario is found in the common resources
A scenario is composed of three parts : selection of the stressers, scheduling of the stress, analyzis of the result
The Stresscloud works on a client-server principle. In a scenario, the server is called registar and we request him the VM needed .
First, we always need a given number of VM to be present BEFORE selecting them. This is for deterministic feature : we want to always select the same VM in the same scenario. So we must always have the same number of VM to start a scenario, if a VM takes longer to start it must not prevent the same scenario from executing again.
Then we can require and assign groups of VM.
see the scenario examples. Some available instructions are stresser-specific, but basically we can do things like
- add a total load to execute , eg send 10MB to a VM or write 1GB to the disk.
- try to execute given number of operations every second.
- synchronize the scenario with time.
- synchronize the VMs with their stresser activity.
todo later