How to find the cause of replication lag on a particular secondary?

A comment worth sharing: at yet another occasion it was confirmed that restarting the mongodb service on the lagging replica helps to remove the lag immediately:

image