Amazon EBS measures volume health using two properties: state and status.
Volume state simply describes the availability of an EBS volume. The possible states are:
The only state that Blue Matador will alert on is the Error state. Volumes in this state are typically unrecoverable, and may be replaced by creating a new volume using a snapshot of the old volume.
EBS status checks run every 5 minutes and determine if the data on an EBS volume is possibly inconsistent. When AWS detects possible data inconsistency on a volume, its status is set to either warning or impaired.
Blue Matador automatically alerts when the volume status becomes either warning or impaired. These alerts must be addressed ASAP to prevent possible data loss or corruption on the affected volumes.
A volume with a warning status may still function at reduced performance. Provisioned IOPS SSD (io1) volumes in particular may enter this state when one is restored from an EBS snapshot. In that case, the warning can be ignored while the volume is being initialized. If a volume is in the warning state, prepare to be able to replace the volume in case it degrades to the impaired state.
When a volume becomes impaired, I/O is automatically disabled on the volume to prevent further data corruption. Check for data consistency on the volume, which can be done using a tool such as fsck. I/O must then be manually enabled on the volume using the AWS console or API. If data consistency is not a concern on a particular volume, you can enable the AutoEnableIO attribute which enables I/O after it has been automatically disabled in the future.