For individual parameterization of virtual users, the GeekLoad has special data sources, feeds and cursors.
The data source directly loads data from different formats before running the test.
The availability of data from these sources on agents is provided automatically.
The data source is integrated into the project script.
Suitable for small data arrays.
The data source is loaded from a file with comma-separated values.
The data source is loaded from the database using the connection string.
Feeds return items from the data source in an order that depends on the type of feed. For each source, you can create any number of feeds, including the same type.
All feeds are functions with one parameter - the cursor name.
Getting a unique value for each virtual user.
Within the same virtual user, the feed will try to use different values, but if they are not enough, then repetitions of values are possible.
If the number of values in the source is less than the number of virtual users, an exception will be thrown.
Getting a random value with a normal distribution from the source. The values can be repeated.
Getting a unique value for each location.
Within the same location, the feed will try to use different values, but if they are not enough, then repetitions of values are possible.
If the number of values in the source is less than the number of locations, an exception will be thrown.
Sequential receipt of the next value within each agent. After the last value is given, the selection starts from the beginning.
In scenarios executed by virtual users, sources and feeds are not available, instead use cursors that get data from feeds.
The use of the cursor is implemented as a function call with the cursor name without parameters.