Hello Zabbix Support Team,

I am running Zabbix Server with about 200 SNMP devices, and for all of them the "Use combined requests"​ checkbox on the host interface is enabled.

After restarting Zabbix Server service, MySQL CPU usage spikes to around 600% (8 core server)​ during the startup phase and remains very high for a while.

I would like to ask for clarification on the following points:
Definition of "Use combined requests"​
Could you confirm the exact internal behavior of this parameter?
My understanding is that it combines multiple OIDs into a single SNMP GET request at the application level, independent of native SNMP bulk/walk. Is this correct, and are there any known side effects under large-scale SNMP environments?
Possible relation to high MySQL CPU​
Given 200 SNMP hosts with combined requests enabled, is it possible that during server restart a burst of combined SNMP polling, history/trend flushing, or cache synchronization leads to heavy MySQL load?
Could this parameter combined with high SNMP concurrency contribute to such a CPU spike (600% on MySQL)?
Impact of disabling this option​
If I disable "Use combined requests"​ (uncheck it for these hosts), will Zabbix fall back to fewer OIDs per request / more sequential polling, and as a result reduce database write/update pressure so that MySQL CPU is expected to drop?
Environment info:
Zabbix version: 7.0.25
Database: MySQL 8.4.10
OS: Redhat8.8
CPU: 8 core
Number of SNMP hosts: ~200
Approximate number of SNMP items total: ~20000
SNMP version used: v2c 
StartSNMPPollers: 8
Any guidance, known issues, or recommended tuning (poller count, cache sizes, DB parameters) would be appreciated.

Best regards,
Alex