After following the various documentation for adding a Windows node, I succeeded in creating a Windows node, but there are several problems:
The Cloud dashboard doesn’t display the graphs in the Cloud Live view.
Metrics are limited to 7 values: cpu, disk, net, os , system, services, export
There are very few metrics, whereas the same windows exporter is used to collect Prometheus/Grafana data, which returns dozens of precise metrics.
The OS isn’t even detected, Netdata declares it as Baremetal even though it’s a VM…etc…
Why does Netdata report so few metrics?
What’s the solution for monitoring a Windows node with all the metrics of the prometheus exporter collects by default?
After a lot of research, it turns out that Netdata isn’t really suitable for monitoring Windows nodes.
It does the job for a few metrics, but it’s doesn’t make sense in using Netdata for this kind of basic metric.
It would be useful if the Netdata team could tell us if it’s possible to get the same metrics as Prometheus/Grafana, given that it’s based on the same exporter.
Could you clarify what metrics you are looking for?
You can find all the metrics we collect on this section Windows | Learn Netdata. This include Server Metrics, AD, Exchange, IIS, SQL Server, .NET Framework.
I already use the Prometheus Exporter for my Grafana dashboard. I collect dozen off metrics : defaults metrics and others like tpc, memory, dfsr with a lot of dashboard
My Grafana dashboard display all these metrics, but the Netdata dashboard only display cpu, disk, net, memory
The issue is not prometheus collector, it’s netdata that does not display multiple dashboards for each metric like Grafana
If you see our docs there we have a bunch of metrics being collected including TCP, if there is some in particular that you see missing could you open a Feature Request using this link ?
Yes, that metdata we don’t collect atm for Windows machines since these are remotely monitored through a Linux machine.
The metdata that you mention is captured by the Netdata Agent when it is running on that machine, for windows the most you can do is set those remote Windows nodes as Virtual Nodes to be seen as Nodes on the dashboard.
We have some use-case setups on our Live Demo environment for Windows, you can see this on this Windows room.
