Steps to reproduce:

1. Have a host with "No SNMP data collection" trigger in PROBLEM state.
2. Move the host to be monitored by a different proxy/group

Result:

1.  SNMP interface status will briefly return 2 (unknown)
2.  Problem is resolved and new problem is generated

Expected:
Original problem is not resolved.

This happens, because the trigger is configured like this:

max(/Alcatel Timetra TiMOS by SNMP/zabbix[host,snmp,available],{$SNMP.TIMEOUT})=0

I don't know if switching host from monitored by server to monitored by proxy group should cause temporary host,snmp,available status 2. I had two unavailable hosts, I switched them to the new proxy group in one mass update, but the "No SNMP data collection" problem was resolved only on one of them.