Running Experiments


Every member is permitted to run the experiments belonging to the team. If you hit the Run button the experiment is immediately started when the following conditions are met:

  1. Experimenting is currently enabled.
  2. There is no other experiment running.
  3. The blast radius resolves to one or more targets.

If one of the condition is not met, you'll get an error message and the execution is not recorded. If all conditions are met the experiment is started.

Experiment Phases

experiment phases

CREATEDThe experiment was created and is waiting for the agents to connect and the state check to be executed.
PREPAREDAll agents taking part in the experiment are connected, the state check reported at least one success and the warm up begins.
ATTACKS_STARTEDThe warm up has ended and attacks were started and are executing right now.
ATTACKS_STOPPEDThe attacks were stopped and the cool down begins.
COMPLETEDAll attacks were executed, the cool down has ended and the state check reported no failures.
CANCELEDThe experiment was canceled by user interaction and all attacks were rolled back.
FAILEDThe execution failed, the reason is indicated in the UI. Can be any one of:
  • Failed attack execution or rollback
  • Failed state check
  • Agent disconnected unexpectedly.

If an agent looses the connection to the platform during an experiment, it will immediately stop and rollback running attacks. There are some attacks (like Stop Container) which can't be rolled back due to it's nature.


Each execution is recorded and the results can be viewed in the Reports section. In the detailed view for the execution you can document any findings you made by the experiment. You can also indicate if any additional action is required to mitigate the incident in the future.

Canceling/Preventing Experiments

In the case the running experiment causes unforeseen severe errors any authenticated user can stop any running experiment by clicking the stop button.

You can prevent further executions if you disable experimenting using the Stop All button. If you disable experimenting, no (manual or scheduled) experiments will be executed until experimenting is continued. Stopping and continuing experimenting can be triggered by any authenticated user.

Need Help? Get in touch with us© steadybit. All rights reserved.