**Is your feature request related to a problem? Please describe.** Agent can be stuck on benchmark's single scenario for whatever reason **Describe the solution you'd like** Add timeout for running single scenario. Think about how to stop/reset simulation?