Amazon EC2 Instance fails continuously, 1/2 status check
Solution 1:
I would monitor several resources of your EC2 instance.
By default, you can only see CPU. If you install the CloudWatch Agent, you can collect additional metrics like RAM or Disk, which the hypervisor can't directly guess.
Then you can review the metrics around the time your instance fails, to check for some unexpected resource consumption.
I've had some issues in the past, where the lack of freeable RAM resulted in a similar behavior as the one you are describing.