Show more VM/LXC stats in cluster overview

[DerDanilo]

It would be really nice to see the stats of the VMs and LXC Containers in der Node and/or cluster overview.

Average/Max/Min usage for all details that are seen per instance:

They could be shown in the overview for example:

On the other hand it would also be nice if the graphs were shown for "selected" VMs/LXC Containers.
It might make sense to only show graps when the user wants to see them.

Otherwise the overview could show mini graps that give an overall overview of how all systems use available ressources over time.

I am not asking for a full stats displaying solution, but to display data in the already quiet nice GUI that was collected already.

Thanks!

 

[dcsapak]

sadly this is easier said than done, because the information in that grid is replicated cluster wide with the pmxcfs and polled every few seconds
so if we would add all the graph data in there, this would have a major impact on cluster operation and site useability

 

[DerDanilo]

That makes sense. Would it be possible to at least give the table some more information to be shown on cluster level?

 

[dcsapak]

in case you did not know: there are several columns hidden by default; they can be shown by clicking the down-arrow besides a column
in case you did already know about that: which information do you miss ?

 

[DerDanilo]

I did not know that. Nice!

in case you did already know about that: which information do you miss ?

Could you add the "average" values to this view as well?

 

[dcsapak]

Could you add the "average" values to this view as well?

what 'average' do you mean? we do not save a single average value for either hour/day/etc..
those values are all in rrd databases and saved for a time interval

but if you want more detailed monitoring, i would recommend to use something built for it, like nagios,incinga or something like that

also we support exporting to exernal metric servers (graphite and influxdb) with that and e.g., grafana you can
build dashboards and overviews as complicated as you wish
(see https://pve.proxmox.com/wiki/External_Metric_Server)

 

[DerDanilo]

Well. Grafana is the long term solution but there is no project out there that provides and updates Grafana dashboards for this.
I found https://grafana.com/dashboards/1147. But who guarantees that this continues working in future?

Currently I have no time to continue supporting this.

 

[dcsapak]

Well. Grafana is the long term solution but there is no project out there that provides and updates Grafana dashboards for this.
I found https://grafana.com/dashboards/1147. But who guarantees that this continues working in future?

what i meant is that you setup your own monitoring and customize it for your needs, i think that any monitoring/graphing solution needs customization/tweaking

but as for our output format, since this comes from the api, it should only change if we change the api (which we seldom do, and mostly in a non-breaking way)

 

[DerDanilo]

Ok, maybe something for the upcoming holidays then.