Alarm didn't work fine from ESXi at NAT environment

[zabx27]

Hello,

I'm using Zabbix 4.0 and performed the alarm test from VMware ESXi.
But,that alarm was not correctly treated at Zabbix and it was not displayed as outage alarm at Zabbix GUI.

#esxcli system snmp test

192.168.0.1 :IP before NAT

10.1.1.1 :IP after NAT

1) Source IP became different that I expected and outage alarm was not displayed at Zabbix GUI.

[root@zabbix001 snmptt]# tail -f snmptt.log

2019/0x/0x 14:57:15 .1.3.6.1.6.3.1.1.5.2 Normal "Status Events" 192.168.0.1 - ZBXTRAP 192.168.0.1 A warmStart trap signifies that the SNMP entity,

2

) Source IP became correct that I expected and alarm was displayed correctly at Zabbix GUI.

[root@zabbix001 snmptt]# tail -f snmptt.log

2019/0x/0x 15:05:31 .1.3.6.1.6.3.1.1.5.2 Normal "Status Events" 192.168.0.1 - ZBXTRAP 10.1.1.1 A warmStart trap signifies that the SNMP entity,

The difference between 1) and 2) was FORMAT ZBXTRAP configuration in "/etc/snmp/snmptt/snmptt.conf".
1) was configured by using FORMAT ZBXTRAP $aA.
2) was configured by using FORMAT ZBXTRAP $ar.

#Same issue didn't happen from another device(iDRAC, SRX) on same NAT environment. I mean that test alarm's source IP became that I expected(after NAT).

Then, I have some questions.
Q1) What's the difference between $aA and $ar ? Why did the issue happen on my NAT environment ?
I checked

http://snmptt.sourceforge.net/docs/snmptt.shtml#DNS but I couldn't understand the difference.

Q2) Why did this change fix the issue ? (It seems that source IP was treated differently between

$aA and $ar.)

Q3) Is there another solution or workaround to fix my issue that I faced ?

Your comment,advise and feedback are highly appreciated so much.