Technical Explanation is included below 

Usually  CELLSRV observes that the service time on the disk increases, its proactively marked as confinedOffline.  This is a first phase of detection of poor performance on the disk. At this stage, we see the alert being generated. After this, the disk is ready for further tests.      Now MS runs the performance tests on the CELLDISK. If it turns out that the disk is performing well, MS will notify CELLSRV that the disk is fine and CELLSRV will notify ASM to put the grid disks ONLINE.   If it comes out that the performance is bad in the tests, the disk status is changed to proactive failure.    Now the GRIDDISK  are dropped from ASM. Then the status of the disk is set to predictive failure.   If MS has indicated that the disk is performing well and hence is set back to normal and a notification for the alert clearance is updated in alert history.   This is all automated operation and no interference is required. 

As the alert is cleared and disk status changed to NORMAL , there is no further action required on this disk.  However in your case as indicated by Travis yesterday this frequently happening, so this need to go to Oracle Support immediately.