Hello,

I use this housekeeping parameter :

Events and alerts

Trigger data : 30d

Internal data : 7d

Network discovery : 10d

Auto-registration data : 10d

In zabbix-server.conf I use :

HousekeepingFrequency=1 (default)

MaxHousekeeperDelete=50000

I view in log, housekeeping work :

  9598:20190704:072400.962 housekeeper [deleted 0 hist/trends, 0 items/triggers, 889 events, 589 problems, 4161 sessions, 0 alarms, 307 audit items in 1.174269 sec, idle for 1 hour(s)]

  9598:20190704:082401.442 executing housekeeper

  9598:20190704:082402.841 housekeeper [deleted 0 hist/trends, 0 items/triggers, 996 events, 608 problems, 4139 sessions, 0 alarms, 309 audit items in 1.375762 sec, idle for 1 hour(s)]

  9598:20190704:092403.327 executing housekeeper

  9598:20190704:092404.534 housekeeper [deleted 0 hist/trends, 0 items/triggers, 868 events, 593 problems, 4140 sessions, 0 alarms, 312 audit items in 1.196285 sec, idle for 1 hour(s)]

I have not CPU overloaded or error on postgresql logs.

I monitoring events table size, it's alway grow up.

For check, I look for find old rows on database (order by clock ASC).

I find old rows on database associate with internal proxy event.

Old event/problem date from more than 30days.

It's not displayed on "Problem screen". (2 years in zoom)

Example 

105 | 3 | 4 | 47974 | 1557151714 | 1 | 0 | 924798637 | No "snmp trapper" processes started. | 0
106 | 3 | 0 | 40051 | 1557151714 | 1 | 0 | 924798637 | Cannot evaluate expression: "Cannot evaluate function "MYPROXY:zabbix[process,snmp trapper,avg,busy].avg(10m)": item is not supported.". | 0
108 | 3 | 0 | 40048 | 1557151719 | 1 | 0 | 926520764 | Cannot evaluate expression: "Cannot evaluate function "MYPROXY:zabbix[process,java poller,avg,busy].avg(10m)": item is not supported.". | 0
107 | 3 | 4 | 47979 | 1557151719 | 1 | 0 | 926520764 | No "java poller" processes started. | 0
109 | 3 | 4 | 47993 | 1557151733 | 1 | 0 | 931154120 | No "vmware collector" processes started. | 0
110 | 3 | 0 | 40059 | 1557151733 | 1 | 0 | 931154120 | Cannot evaluate expression: "Cannot evaluate function "MYPROXY:zabbix[process,vmware collector,avg,busy].avg(10m)": item is not supported.". | 0
111 | 3 | 4 | 47994 | 1557151734 | 1 | 0 | 931534924 | No "vmware collector" processes started. | 0
112 | 3 | 0 | 40058 | 1557151734 | 1 | 0 | 931534924 | Cannot evaluate expression: "Cannot evaluate function "MYPROXY:zabbix[vmware,buffer,pused].max(10m)": item is not supported.". | 0
178 | 3 | 4 | 48119 | 1557825959 | 1 | 0 | 588958251 | No "vmware collector" processes started. | 0
179 | 3 | 0 | 40179 | 1557825959 | 1 | 0 | 588958251 | Cannot evaluate expression: "Cannot evaluate function "MYPROXY:zabbix[process,vmware collector,avg,busy].avg(10m)": item is not supported.". | 0
181 | 3 | 0 | 40178 | 1557825960 | 1 | 0 | 589230206 | Cannot evaluate expression: "Cannot evaluate function "MYPROXY:zabbix[vmware,buffer,pused].max(10m)": item is not supported.". | 0
180 | 3 | 4 | 48120 | 1557825960 | 1 | 0 | 589230206 | No "vmware collector" processes started. | 0
183 | 3 | 4 | 48100 | 1557826000 | 1 | 0 | 596832807 | No "snmp trapper" processes started.

 On another environnement, same thing, I have some row from "ghost" events from custom trigger (not only internal event).

How to clean up this old row ? It's cause by not closing problem ? how to close problem if not display in web interface ? Change status with query ?

Regards,