happily using a single docker node to monitor many Windows servers as recommended, with windows_exporter.exe
i setup wmi.conf and prometheus.conf and get all my servers duped with different data sets, and some redundancy
should i stick with either one or both?
what is the best practice? i am afraid this may double the number of requests to each windows agent
I would suggest using wmi. Both will do, but using prometheus you get less meaningful charts due its generic nature. And we have predefined alarms for wmi.
i am afraid this may double the number of requests to each windows agent
You are right.
thank you, will stick with wmi for now!
only drawback is that i don;t see the processes and services i monitor… below is the command i use for the exporter:
msiexec /i windows_exporter.msi ENABLED_COLLECTORS="cpu,cs,logical_disk,net,os,service,system,memory,process" EXTRA_FLAGS="--collector.service.services-where=""Name LIKE 'MyService%%'"" --collector.process.whitelist=""MyProcess1.+|MyProcess2.+"""
@audioscavenger current wmi implementation collects only (ENABLED_COLLECTORS):
Back then i added the most common (from my point of view) metrics. I update wmi if there is a request.
If you are interested in any other metrics i suggest to open a feature request. I can add them relatively quickly.