After an outage of Flow, you may see considerably high CPU usage by Rules across multiple processes.
When Flow is down, Agent instances will queue up data to send through Flow when it is once again available. If this downtime extends for several days, it is possible that you will have a considerable amount of data queued up on the Agent side.
When Flow becomes available again, all of this data will be pushed to Flow and be evaluated against Rules for any potential violations, which results in a higher than normal load on Rules thus the higher than normal CPU load.
In the event that this occurs, you should see the CPU usage of Rules return to a normal level once processing completes.