EC2 RebootInstances performs dirty restart

Hello, I am using eventbride to issue EC2 RebootInstances API Call to a number of machines. Most shutdown and boot well. 1 does not seem to receive the shutdown request and does a dirty restart.

I can see that successful machines have an event 109: 'The kernel power manager has initiated a shutdown transition. Shutdown Reason: Kernel API' . The failed machine does not.

Checking the SSM logs on a good and bad machine, the log before the reboot looks identical other than time stamps. These two comparison machines are same ssm agent version, same windows, instance type.

Cloudtrail / eventbridge logs don't show failures.

Can anybody give me a pointer where to look for this issue? thank you