Ask HN: Do you try to understand the root cause for already solved problems?

I just solved an authentication issue between Git and Jenkins using a lot of trial and error, semi random advice from SO and rolling back some code. It works now but I still don't fully understand why it works.

This is not uncommon in devops and test, but happened to me more than once in embedded code and even in other places. Sometimes I spend some official or private time investigating it, sometimes I put it into a backlog and in some cases I don't even bother due to lack of time.

How much time, if any, do you spend investigating something that is already working? Will you loose sleep over it?