Distributors main page

• General information about interface: Distributor
• Available implementations: Implementations of distributors
• How to customize: Custom distributors
• How to run test with randomized order of requests: Access SUT with random order of endpoint / query pairs
  
  

General info

Distributors provide pairs of endpoints and queries for invokers
Before start of the test distributor is combining all endpoints and queries according to user setup and stores these combination in internal list.
Important: mentioned list is shared by all threads that produce load. it is not possible to have separate list per workload thread
Before every invoke distributor is providing single pair of endpoint / query to invoker. Test run sequence you can find here: Test execution sequence

Example of distributor setup in XML:

Following XML code should be included in test-description section

    <!-- begin: following section is used for docu generation - invoker usage -->
    <!--Description of single scenario: "Visit page"-->
    <!--This scenario is used as base element and later will be inherited by other elements in this file-->
    <scenario id="pageScenario" xsi:type="scenario-query-pool">

        <!--distributor - how to combine endpoint and query-->
        <query-distributor xsi:type="query-distributor-round-robin"/>

        <!--invoker - how to send information to endpoint-->
        <!--here is example of custom invoker usage-->
        <invoker xsi:type="invoker-class" class="${package}.invoker.PageVisitorInvoker"/>
    </scenario>
    <!-- end: following section is used for docu generation - invoker usage -->

Variants of distributors available in XML:

query-distributor - documentation of distributor element in XML schema
queryDistributorAbstract - types of distributors available in XML schema. See 'Sub Types' section of man page
How distributors mentioned above are implemented you can see in section: Implementations of distributors

Access SUT with random order of endpoint / query pairs

Pairs of endpoint / query are collected into list single time during Jagger start up.
By default all threads are accessing elements of mentioned list in the same order: from first element to the last, doing this in cycle.
If it is necessary to execute endpoint / query pairs in random order, set randomSeed attribute of f.e. query-distributor-round-robin equal to some integer value

Access SUT with random order of endpoint / query pairs