The IRecoveryClient shall be modified so that LaunchManager-internal HealthMonitor only communicates supervision failure. The recovery state shall be taken from Launch Manager ("configuration_manager_.getNameOfRecoveryState"). There is no need for a different recovery state in case of supervision failure vs. process crash.
Furthermore, the independent timeout monitoring that is currently done as part of the RecoveryAction can be removed
As the Watchdog will in future not be servicing by the LaunchManagers HealthMonitor Thread, we shall introduce an error flag into the IRecoveryClient interface. In case of error the flag will be set and can be retrieved by the LaunchManager.