Replayer + Tabjolt

Version 2

    Tabjolt is a performance and capacity planning tool that Tableau released as community supported tool few years ago, it is available on Tabjolt github, it is one of the widely used tool internally at tableau as well as external customers. One can just point Tabjolt to a Tableau server and get started with using it, it collects performance counters from tableau server, tool allows user to set number of user simulation.

     

    While tabjolt is good for scenarios where capacity planning is needed for new deployment, for organizations which already have established tableau server with regular user access Replayer gives advantage of using your existing traffic pattern and scaling up or down based as per the need for capacity planning. To take advantage of Tabjolt’s performance counter collection we have integrated Replayer into Tabjolt, it also allows saving results in DB which makes it easy for results comparison. Tabjolt is used as the command line tool to start of a Replay test by specifying the ReplayFile containing real customer usage pattern.

     

    After creating a Replay file using logshark and filtering to interesting time interval use tabjolt to run the test.

     

    Running Replayer using Tabjolt

    Eg: go.bat --t=testplans\Replay.jmx  --d=5 --c=1 --u=http://tableauserver --s=false --note=Replaytrial --r=description --w=NA --z={replayFile:Playback_25_09_-16-43-43.json,refreshCache:no}

     

    select test with --t option to specify the Replay testplan for Tabjolt (replay.jmx).

    --t=testplans\replay.jmx

     

    Use the --u option to specify the URL of the Tableau Server.

     

    To pass the Replay command parameters to Tabjolt, you use the --z= option. The first parameter is replayFile that specifies the Replay playback file, it is necessary to copy the replay file to the running directory..

    --z=replayFile:ReplayFile.json

    To specify additional options, use a comma (,) and the option=value pair. For example, the following example includes the replayfile,

    multiuserAuth, refreshCache settings

     

    -z={replayFile:Replay-interaction.json,,refreshCache:no;multiuserAuth:yes}

    for multiuserAuth there should be a userpool.csv in the config folder which Replayer will use for picking up users for authentication.

     

    --d=1

    Sets the Tabjolt duration option. Set this option to 1. The actual duration is determined by the Replay file.

     

    --c=1

    Sets the Tabjolt threadCount option. Set this to 1.

     

     

     

    For capacity planning traffic pattern can be scaled up or down with traffic multiplier using Replayer UI->Alter Traffic->Multiply, values can be something like .5, 1.5, 2, 5 etc. Replayer will create a new Replay file with the scaled traffic pattern that can be used for capacity planning.

     

    Results analysis

    For viewing the results after the run using tabjolt one can use the Tabjolt viz and compare the results that are recorded in the DB. This can be used in addition to vizzes that comes with Replayer that uses results from csv.