Problem Summary: On heavy workload right tikv pods are occasionally getting rebooted due to out of memory.
Prior to version 3.1, TiKV block-cache parameter is not set automatically, and might cause OOM on heavy load.
# shared = true
# Normally it should be tuned to 30%-50% of `tikv.resources.limits.memory`, for example: 32Gi -> 16GB
# capacity = "1GB"
Starting from 3.1 is automatically set by default to 50% of total memory