Cockroach db pod continue increase memory graph

Guys, I am using Cockroach db on GCP Kubernetes. There is running 3 pods of cockroach db. I am observice that it’s restarts randomly and due to that my application thrown an error while pod restart. check the below pod screenshot. What could be reason of this pod restart ?

image

Also I am observing that memory only increased of that pods. It should be also decrease when load is less on database. I have configured --cache 750Mi --max-sql-memory 750Mi in cockroach db kuebrnetes configuration. You can see in below screenshot that once pod restarts it going up and increase memory continually, Can someone explain me that how memory usage works.