Please consider the following feature request involving making better use of typeperf information on windows agents :

1) In the Windows agent, build in the ability to collect the results of "typeperf -qx" and pass that list back to the Zabbix server.

2) On the Zabbix server, in the web interface under a particular host, allow the admin to view this typeperf list.

That alone would centralize this very useful information so people don't have to check typeperf on individual machines prior to adding perf_counter Items.

Once that is in place, it can be taken a step further :

3) While viewing the list of typeperf results for a host in the web GUI, allow a number of individual items to be checked off, and then a button could be clicked to automatically add perf_counter Items for each checked off typeperf line. Default values could be used for many of the fields, which could then be individually edited later if desired.

This would allow fast creation of multiple perf_counter items, and since they would be based on typeperf collected right from the target machine, they should all work properly on that Windows host. It would also eliminate the manual entry of all those long string definitions and reduce the chance of typos.

Once this functionality is is place, it would be useful to be able to create a template based on that host. ( see ZBXNEXT-595 ).

For future thought: Once the concept of an agent collecting a list of available monitor targets and passing that list back to the Zabbix server for selection is in place, something similar could be done for other monitoring targets such as "netstat -l" (listening ports), available SNMP monitors, etc. These would also be applicable to Linux hosts.