I've seen an issue where cloudstack would stop some VMs after they have
been started. Posting the solution for whoever comes across this problem.
I've noticed following pattern:
VM was trying to be migrated to another host while another thread would
stop the VM within 10 minutes of the start time.
The solution was to look for a message in the log that has
"(981,VmWorkJobQueue, 2371) is reaching concurrency limit 1" continuously.
Delete the reference id of "2371" from cloud.sync_queue_item
Delete the reference id of "981" from cloud.sync_queue
