I updated zabbix from version 1.6.7 to 1.8.1 one month ago. Since of then, I realize an increasing memory consume by zabbix server process (see screenshot 1). As a result, I have to restart zabbix server process two or three times a day. I've looked for similar problem on the Internet and I found only one thread that made sense: http://www.zabbix.com/forum/showthread.php?t=7126.
Let me now describe some aspects of my zabbix configuration:
Status of Zabbix
Parameter Value Details
Zabbix server is running Yes -
Number of hosts (monitored/not monitored/templates) 258 185 / 8 / 65
Number of items (monitored/disabled/not supported) 9956 1817 / 8092 / 47
Number of triggers (enabled/disabled)[true/unknown/false] 1003 978 / 25 [73 / 0 / 905]
Number of users (online) 23 1
Required server performance, new values per second 20 -

I've compiled zabbix with the following options: ./configure --prefix=/usr/local/zabbix --with-pgsql --with-ldap --enable-server --enable-agent --with-jabber --with-net-snmp --with-libcurl
Average CPU utilization of zabbix server machine is less than 5%.
Average CPU utilization of database machine is less than 4%.
There are no web monitoring and agent zabbix active checks enabled.
There are several snmp, calculated and passive check based items enabled.
After I upgrade zabbix to version 1.8.1 I was using mysql database and the same memory problem was observed.

Follow now some of softwares installed on my CentOS box (where zabbix server is installed). See screenshot 2 for some other details:
postgresql-8.1.18-2.el5_4.1 | postgresql-devel-8.1.18-2.el5_4.1 | postgresql-libs-8.1.18-2.el5_4.1 | postgresql-devel-8.1.18-2.el5_4.1
curl-7.15.5-2.1.el5_3.5 | curl-devel-7.15.5-2.1.el5_3.5
php-5.3.1-1.el5.remi | php-pgsql-5.3.1-1.el5.remi | php-gd-5.3.1-1.el5.remi | php-common-5.3.1-1.el5.remi | php-pdo-5.3.1-1.el5.remi| php-ldap-5.3.1-1.el5.remi | php-bcmath-5.3.1-1.el5.remi | php-xml-5.3.1-1.el5.remi | php-cli-5.3.1-1.el5.remi |php-mbstring-5.3.1-1.el5.remi
iksemel-1.2-13.el5 | iksemel-devel-1.2-13.el5
libxml2-2.6.26-2.1.2.8 | libxml2-devel-2.6.26-2.1.2.8 | libxml2-python-2.6.26-2.1.2.8
openldap-2.3.43-3.el5 | openldap-devel-2.3.43-3.el5
net-snmp.x86_64 1:5.3.2.2-7.el5_4.2 | net-snmp-devel.i386 1:5.3.2.2-7.el5_4.2 | net-snmp-libs.i386 1:5.3.2.2-7.el5_4.2 | net-snmp-utils.x86_64 1:5.3.2.2-7.el5_4.2

Considering the thread mentioned before, I've started to monitor my snmp pollers and they are probably the responsible for the excessive memory consume. See screenshots 3, 4 and 5 for some measures that prove my suspects. As suggested by the thread, I've update my snmp software by compiling net-snmp 5.5 sources (the last binaries version available for CentOS is 5.3.2). Then I recompiled zabbix and installed new generated binaries (server and agent). Unfortunately, the memory leak problem persists. Is this a known problem on CentOS? Has anyone faced the same problem?