Hi,
I have a dozen of server to monitor. I want to compute usage statistics ( nb user, cpu, RAM, GPU, in time with a given threshold for minimal activity) at a given resolution, aggregated over the pool of server and in daily or montly reports (to have macroscopic stat per period , per server or per whole pool. I begin with Netdata, is possible to do it , quite simply?
For that time, i look at snapshot and custom plugins…
Hermann
Hello @hermann74 and welcome in our community
We’re getting a little bit out of the monitoring/ troubleshooting areas with this request. For sure the Netdata Agent can fetch you all the metrics you need and even more than the ones you asked for. But to perform analysis on these data, you must export them to an another service/app/db and then perform your analysis/ create the reports you want.
Before starting any effort I would suggest you to look into the following docs:
Thanks @Tasos_Katsoulas for your help and suggestions. I read carefully these topics.
If I understood well , i should “send metrics to another time series database for long-term storage or further analysis using Netdata’s exporting engine.”
In the whole list of supported database “Exporting reference | Learn Netdata”, I would choose:
- graphite (a document here : Export and visualize Netdata metrics in Graphite | Learn Netdata )
- OpenTSDB : an SGBD for times series
- Prometheus : “We can then use Grafana pointed at Prometheus to obtain long term metrics Netdata offers”
According to you, what is the simplest to deploy for my purpose ?
Not an expert on data analysis. (Personal opinion) I would go with OpenTSDB and R/python for analysis. But this question also depends on the technologies you are familiar with.