I have some powershell scripts scheduled on a windows server and want to track their memory and cpu utilization. And I do have a working solution, but I just wondered what's the best practice here.
So I created a WMI input for Win32_process, which has the "CommandLine" for each powershell.exe instance, and that includes the script name.
But out of the box, Splunk cuts off that CommandLine field at the first space.
I guess my question is, if there are any props.conf options (or in any other configuration) to let it read that full "line" and not stop at the spaces. It seems to be clever enough to read the WMI feed in a way where it puts each field=value into one line in the _raw data.
The server in question has a universal forwarder installed. And that has an app with the wmi.conf in it:
[WMI:powershell.exe]
index = perfmon
disabled = 0
interval = 60
wql = SELECT CommandLine, ProcessId, WorkingSetSize, KernelModeTime, UserModeTime from Win32_process WHERE Name = 'powershell.exe'
And on the indexer, the sourcetype is configured in the props.conf like this:
[WMI:powershell.exe]
FIELDALIAS-dest_for_perfmon = host AS dest
FIELDALIAS-src_for_perfmon = host AS src
LINE_BREAKER = ([\r\n]+)
NO_BINARY_CHECK = true
category = Custom
disabled = false
In the end, my goal is to get that script name, so I just added a field extraction for that
EXTRACT-ScriptName = CommandLine=powershell.exe+\s+\-command+\s+"&\s+'*(?P<ScriptName>[^("|')]+)
Is that the best way of doing it?
