Steps to reproduce:

1. After upgrade from 5.0.28 to 6.0.20 Cisco and APC devices CPU overload occurs on devices one by one.

Result:
See screenshot...

The device is monitored normally for a long time, 30+ days.

At random times, the CPU load increases to 90%. Monitoring configuration was not changed at the time of or before the problem.

Performance analysis on the device shows a high load of SNMP processes.

TCP dump on the Zabbix server shows about 60,000 packets per minute to the problematic device, while on devices where the problem is not observed there are about 100 packets per minute.

Disabling device monitoring does not solve the problem. Restarting the Zabbix server solves the problem.

Cisco switch CPU load grap attached.

Cisco CPU load details attached.

Wireshard screenshot attached. You can see there that the number of requests is off the charts.