After restarting postgresql I get "[Z3005] query failed: [0] PGRES_FATAL_ERROR:FATAL:" errors in zabbix log. They stabely appear from time to time, until I restart zabbix-server.
Errors differ slightly on the basis of "DBHost" configuration (remote or localhost). See logs for details.
Don't know, is this critical or not?

To reproduce:
1) set DEBUG to 3 in zabbix_server.conf
2) run DB service, run zabbix server, check that they work
3) stop DB service, wait 10-20 sec, start it again
4) see zabbix logs, PGRES_FATAL_ERROR will appear periodically (think, it depends on the count of hosts and items)