While analyzing zabbix_server -R diaginfo, identified three item IDs dominating the history write cache. Found that all three corresponded to the official vmware.eventlog[url,skip] key, assigned to three different VMware hosts.

• collected value counts were 1.2M, 600K, and 400K
• it ocurred within less than an hour of Zabbix server uptime
• impacted performance and agents availability

Fix:
Disabled the item at the template level, preventing further event log collection.
Within minutes, observed a significant improvement in history syncers' performance

The issue is that if someone who is unaware of that downloads the updated template and set up vmware monitoring, can crash the entire infrastructure.