Starting from v5.1, you can designate different data to different load generators. For example, an application accessed from different geographical locations can have different credentials in each location.
In this case, you can create a test with multiple load generators (agents) in different geographical locations. Then create a dataset with a set of credentials, and each credential can have a designation which load generator will consume it.
To do this, do the following:
- Create a credentials dataset
- Add a column called "Agent"
- Fill in the data.
- In the "Agent" column, specify the Agent name where the credential row will apply.
If the Agent cell is left blank, then the row will be consumed by all agents.
- Set the Agent-binding column property to "Agent"
What will actually happen during the test run is both Load generator NewYorkAgent and SydneyAgent will have the credentials dataset. However, the data will be different.
- Load generator NewYorkAgent will have a credentials dataset with the first 3 rows of the original (username1-3)
- Load generator SydneyAgent will have a credentials dataset with the last 4 rows of the original (username4-7)
Automatically designate data to agents
The above section described how to assign every row a designated load agent. However, if exact row designation is not critical, StresStimulus can automatically designate every row based on the load agents' mix weights.
To enable automatic dataset row destination, set the Distribute dataset rows to load agents automatically property to Yes.
Suppose a test with the following elements:
- A credentials dataset with 100 rows
- 3 load agents with the following mix weights:
- LA1 mix weight 1
- LA2 mix weight 2
- LA3 mix weight 1
The rows will be distributed as follows:
|Load Agent (mix weight)||Rows|
- No labels