Currently we are monitoring the number of returned, fetched, inserted, deleted and updated rows:
It can also be helpful to monitor the number of queries. Maybe postgres even provides statistics of cache hits/misses.
Monitoring the number of SQL queries can help analyzing performance, and also possibly find accidental code changes, e.g. a removed
prefetch in DBIx::Class (like this one: https://github.com/os-autoinst/openQA/pull/3969)