In general container crash should be related with crash / unavailability of the process that is working inside. So based on this you can turn on alerting In process group about process crashed. This should work. About CPU, there is option for defining custom alerts in Dynatrace for processes as well. From container level it may be possible for defining threshold for proper metric (you need dt version 1.72 with new metrics framework) but I'm not sure if this will work for built in metrics (this is in general for plugins and custom metrics I think).