Execution Settings

Execution objects represent actual underlying tool executions. You can launch unlimited number of JMeter's, Gatling Tool's, Grinder Tools, etc. Executions are configured under top-level config key execution. Specifying single execution config is equivalent to specifying array of executions with single element, for example:

  scenario: scenario_name    

is equivalent for

- scenario: scenario_name

However, users are encouraged to use array notation always to leverage the arrays auto-join capability when combining multiple config files into one. See config merge rules for more details on this.

There are load profile and scenario settings that are common for all execution types, and each executor type can also have its own settings.

Executor Types

Taurus tool may use different underlying tools as executors for scenarios. Currently supported tools are:

Default executor is jmeter and can be changed under general settings section.

  default-executor: jmeter

You may contribute your efforts in supporting requests-scenarios for your favorite tool by discussing this on project forum.

Load Profile

Execution has several options to set load profile settings. Support for options is specific to executor type. Available settings are:

  • concurrency - number of target concurrent virtual users
  • ramp-up - ramp-up time to reach target concurrency
  • hold-for - time to hold target concurrency
  • iterations - limit scenario iterations number
  • throughput - apply RPS shaper, limiting maximum RPS to throughput, requires ramp-up and/or hold-for
  • steps - allows users to apply stepping ramp-up for concurrency and rps, requires ramp-up
  • scenario - name of scenario that described in scenarios part (see below)
- concurrency: 10
  ramp-up: 15s
  hold-for: 2m
  iterations: 1000
  throughput: 20
  scenario: scenario_name


Scenario is a sequence of steps that is used to build script for underlying tool (e.g. generate JMX file for JMeter). It is described in special scenarios top-level config element. There are three examples of scenario syntax:

  get-requests:                     # normal form: scenario is dictionary
      - http://localhost/1
      - http://localhost/2  
  only-script: grinder_script.py    # short form: just script

- concurrency: 10
  hold-for: 1m
  scenario: get-requests  
- executor: gatling
  concurrency: 5
  iterations: 10
  scenario: only-script
- hold-for: 20s
  scenario: my_jmx_file.jmx         # shortest form: only script file name  

Startup Delay

You can run different executions at different times with delay option:

- concurrency: 10
  hold-for: 20s
  scenario: main
- concurrency: 20
  hold-for: 15s
  scenario: main
  delay: 10s

    - http://blazedemo.com/

By this way, the first execution works 10 seconds, then two executions will work 10 seconds together, then the first will stop and the second will complete its work in 5 seconds.

Another way to schedule is usage of start-at:

- concurrency: 10
  hold-for: 20s
  start-at: '2020-03-25 23:15'  # must be string
  scenario: sample
    - http://blazedemo.com/

Supported time formats are:

  • HH:MM:SS
  • HH:MM

Additional Files

When your execution requires additional files (e.g. JARs, certificates etc.) and you plan to send tests to the –°loud, you may use files option of execution and list paths for files there.

Sequential Execution

By default, Taurus runs items under execution in parallel. To switch it into sequential mode, run it with -sequential command-line option. This is an alias for this setting:

    sequential: true

Keep in mind: sequential execution doesn't work in the –°loud mode. And as modules start sequentially, Startup Delay doesn't matter in this case.