Have you tried turning up the retries for the health check? My thinking is that if it's determined that it's not a resource issue, host memory, cpu, external storage volumes, etc., then the cause of the container reboot is most likely failing the health check.