Zabbix HA & DR Solutions??

[tsteel]

I inherited a Zabbix 5 environment that supports 7000 nodes. There are two Zabbix servers (one in each datacenter). Daily the database is backed up and transferred to the stand-by server and imported which takes about 12 hours.
If the Zabbix server losses connectivity to the network or the database becomes extremely busy it generates 1000s of ticket
I need to remove single points of failure and make this HA so I was leaning towards rebuilding the environment using Zabbix 6.x, redundant DB, Zabbix, & Frontend servers.
First question, is this doable across datacenters (DR) or am I tempting fate by causing a split brain if the datacenters loss connectivity?
Or am I better off building out in the main datacenter and just putting a proxy in the smaller datacenter?
Does anyone know of a best-practice architectural document that might cover this?
Thanks

[Amiru]

Hi, I'm also looking for the same answer, as far as I understand, the native HA in zabbix 6 doesn't cover redundant DB right?
what i have in mind is native zabbix 6 ha using master-slave db in separate datacenter which i have to manually change the database pointing when needed.
I haven't test it yet, just finished migrating my zabbix 4 to 6.0 2 days ago.

[cyber]

zabbix native HA covers only server processes.
if you build HA DB, you do not have to switch over manually, there a re tools for that, pg_auto_failover for example for PG, there are others (pg-pool etc). MySQL has also their clustering options.
You can have additional haproxy in front of db cluster, which checks which one is master and redirects queries there...