Based on https://www.rfc-editor.org/rfc/rfc3414#page-14, section 2.2 Replay Protection, the engineBoots should increase every time the snmpd is restarted and per device reboot. The issue is that VyOS is not persisting this engineBoots counter across reboots, causing issues with the SNMP sessions with the Servers. After the reboot the engineBoots value returns to 1, causing existing sessions with SNMP managers (Servers) consider this snmpv3 trap packets old, and discarding them. It gives logs like this (Log from Zabbix Server):

usm: Verification succeeded.
usm: Message too old.
reboot count invalid

Expected behavior: Even after reboot, the VyOS should persist the old counter engineBoot value, as other Linux distributions do. We have tested with Debian Linux, and this engineBoot counter persists after it.

This has been tested in 1.4.x and 1.5.0 VyOS version.

Workaround:
+ By just increasing the engineBoot manually to a higher value than the SNMP manager server already has fix it temporally, but after a new reboot issues occurs again.

EngineBoots (VyOS) >= EngineBoots SNMP (on the Server cache), otherwise the SNMP messages won't be accepted.