Steps to reproduce:

thereare 2 physical F5 machiens with high specs (around 99,000+ OIDs). when are adding it into zabbix monitoring, we can see frequent tiemouts.
but when we are doing commandline snmpwalk to the f5 machine (UDP with timeout 30sec, and snmpwalk via tcp, its working fine)

Result:
See screenshot...
See log file...
See memory dump...
Expected:
See screenshot....
See attached patch file...