Build a scalable and highly available Zabbix monitoring system by scaling it on database level with MySQL InnoDB cluster. It guarantees data consistency and automated conflict resolution, automated failover, and self-healing.
There are a few ways one can gather custom metrics with Zabbix, for example, via user parameters or Zabbix sender protocol. While either method can work for simple metrics, creating a more wholesome approach (with data collectors, persistent connections etc) and interfacing it with Zabbix may not be as trivial.