It saves the system state into an instance of class Log-System-State-Cluster in data.pr_log_system_state daily at 1am. You can then retrieve your system state for the current day or for past days from Admin Studio. See the following for more information:
Node state is saved by pyPersistNodeState on node start up. Cluster state is saved once a day by the pyPersistClusterState job scheduler. Cluster state data is saved for 90 days and is purged by the pzClusterAndDBCleaner job scheduler.