be able to tell when something is working and when it's not working

monitoring is fundamental to continuous delivery. For a given system you should be able to tell, via application logs and alerts (alerts should be relevant), when things are progressing as normal.

Knowing what is normal simplifies determining what is abnormal.