Hello everyone,
first post, fresh member so I'll try not to make a fool of myself.
Current topology for Zabbix monitoring system:
- Zabbix server in company HQ
- Zabbix proxies installed in every company site (6 proxies, one on each site, located worldwide)
- Zabbix agents deployed on every Linux / Windows server (a bit over 350 Linux and Windows servers)
- network equipment monitored through SNMP (a bit over 200 network components)
- various network terminals / appliances monitored also through snmp (around 450-500 other devices)
Each agent connects to the regional proxy which connects to the main Zabbix server in the company HQ.
What we try to accomplish now:
- we deployed a secondary Zabbix server in the most distant company site;
- we would like to configure a HA cluster with a conditional alerting system
What I understand from HA cluster with conditional alerting:
- a HA Zabbix servers cluster (either using HA proxy, or some internal clustering system (I don't really know if Zabbix has that in place);
- a PSQL HA cluster for the PSQL servers on Zabbix servers;
- all proxies send the same information to both Zabbix servers;
- when the main Zabbix is online and functioning properly, ONLY the main server sends the alerts for problems;
- when the main Zabbix server stopped working for any reason, or if there is a PSQL problem, the secondary server takes over everything, including alerts.
So, basically, I think it can be summed up like this:
- we need a HA Zabbix cluster (active-passive);
- cluster needs to use 2 different PSQL servers also clustered;
- failover needs to be done if either main Zabbix or PSQL server has a problem and the full role needs to be taken over by the secondary server;
- in fully working state, the secondary server only receives the data, but keeps the alerts dormant as they are managed by the main server.
Thank you
Dan
first post, fresh member so I'll try not to make a fool of myself.
Current topology for Zabbix monitoring system:
- Zabbix server in company HQ
- Zabbix proxies installed in every company site (6 proxies, one on each site, located worldwide)
- Zabbix agents deployed on every Linux / Windows server (a bit over 350 Linux and Windows servers)
- network equipment monitored through SNMP (a bit over 200 network components)
- various network terminals / appliances monitored also through snmp (around 450-500 other devices)
Each agent connects to the regional proxy which connects to the main Zabbix server in the company HQ.
What we try to accomplish now:
- we deployed a secondary Zabbix server in the most distant company site;
- we would like to configure a HA cluster with a conditional alerting system
What I understand from HA cluster with conditional alerting:
- a HA Zabbix servers cluster (either using HA proxy, or some internal clustering system (I don't really know if Zabbix has that in place);
- a PSQL HA cluster for the PSQL servers on Zabbix servers;
- all proxies send the same information to both Zabbix servers;
- when the main Zabbix is online and functioning properly, ONLY the main server sends the alerts for problems;
- when the main Zabbix server stopped working for any reason, or if there is a PSQL problem, the secondary server takes over everything, including alerts.
So, basically, I think it can be summed up like this:
- we need a HA Zabbix cluster (active-passive);
- cluster needs to use 2 different PSQL servers also clustered;
- failover needs to be done if either main Zabbix or PSQL server has a problem and the full role needs to be taken over by the secondary server;
- in fully working state, the secondary server only receives the data, but keeps the alerts dormant as they are managed by the main server.
Thank you
Dan