[pve-devel] Extra Node Summary Information

Alexandre DERUMIER aderumier at odiso.com
Fri Sep 11 16:10:00 CEST 2015


>>The first stage of what I want to do is make a way (possibly a plugin)  that enables Proxmox to collect a variable number of metrics per node so that it could be stored in a RRD or sent to something like Graphite.  Assuming my understanding to Graphite and influxdb are correct.
 
yes.

you can also use collectd, a local daemon which can send host datas to graphite,influxdb,....
They are a lot a collectd plugins
https://collectd.org/wiki/index.php/Table_of_Plugins

then you use grafana (www.grafana.org), to make dashboard using metrics from graphite/influxdb. (and mix results from collectd metrics, and proxmox stats metrics)


----- Mail original -----
De: "Mark Davis" <smiley at smileyshouse.net>
À: "Thomas Lamprecht" <t.lamprecht at proxmox.com>, "pve-devel" <pve-devel at pve.proxmox.com>
Envoyé: Vendredi 11 Septembre 2015 16:04:33
Objet: Re: [pve-devel] Extra Node Summary Information

Extra Node Summary Information 


Thank you for your response. 



I have already setup a VM and been able to compile stable-3 packages. Still trying to get my head around the commands for git though, but I guess that will come with time. 



I thought that Graphite and influxdb was for centralised collection of the metrics and that Proxmox could only send the metrics that it currently collects. The first stage of what I want to do is make a way (possibly a plugin) that enables Proxmox to collect a variable number of metrics per node so that it could be stored in a RRD or sent to something like Graphite. Assuming my understanding to Graphite and influxdb are correct. 



Searching on the net, a lot of people have asked about adding other monitoring metrics. I understand that every computer is different and that the current metrics are the ones guaranteed to be measured the same way on every piece of hardware. Just wanted to know if the dev team was receptive to what I want to try and develop as a possible feature. Not expecting any guarantees, because you guys don’t know me from a bar of soap, but I just don’t see the point of doing a whole lot of work if the submission is not even going to be considered. Also, I would have thought it a bit arrogant of myself if I submitted anything without first conversing with the established development team. 



I would like to say to anyone that has worked on Proxmox, Thank you. I love it and am constantly pleasantly surprised by the added features each new version and how things just seem to work. 



Mark 






From: pve-devel [mailto:pve-devel-bounces at pve.proxmox.com] On Behalf Of Thomas Lamprecht 
Sent: Friday, 11 September 2015 5:02 PM 
To: pve-devel at pve.proxmox.com 
Subject: Re: [pve-devel] Extra Node Summary Information 




Hi, 

I appreciate your interest in PVE and the open source development. 

To point it out first http://pve.proxmox.com/wiki/Developer_Documentation may help you somewhat. 
While it may look like a lot at the beginning after you initial setup a machine (maybe an VM with nested Proxmox) building the packages and changing something becomes easier, if your already have basic development skills it helps you much. 

As you can read also in the last paragraph we need an contributor agreement (it ensures that you hold the copyright of your work when you contribute to the project, but we may use it for the project, legal stuff). 
Contributing (especially to the first) to Open-source projects may have, unfortunately, some initial troubles and a learning curve for the standard tools (like git), don't let you scare from that. 

To be direct: The RRD stuff you proposed probably won't get into the PVE upstream code. 
Although it's for sure a nice and cool feature there are already tools out there which are much better for the job. PVE is more a cluster framework, monitoring should be done by monitoring frameworks. 

But we recently applied an new status plugin, so you could use https://influxdb.com/ or also Graphite to gather all the information, see nice graphs and timelines. 

But if you say, hey I do it my way, still post your changes (as a patch), but don't be disappointed if they get rejected. :) 

Note, take this answer as an personal from me, not Proxmox or affiliates. 

Best Regards, 
Thomas 


On 09/11/2015 07:53 AM, Mark Davis wrote: 




Hi there, 

I have never had anything to do with an open source project before, so if I am doing something in the worng place, please can somebody point me in the right direction. 

Fow a while now I have been using ipmitool on some Proxmox boxes to collect temp and fan info and with my own service logging it to an rrd db. With a few small changes to the proxmox files (which I have to change after an upgrade) I can see graphs in the interface. 

This works well on a single node setup, but in a cluster obviously the status data does not get propergated to all nodes. 

Whomever is in charge of what features are added, I would like to try and add the ability for some extra status information to be stored in an additional rdd db for each node with an editable script defining how the data is collected. Hopefully with diffent data being colledted per node and eventually have it be able to be setup through the proxmox interface. 

I understand that I will have dicuss this in much greater detail. I do not want step on anybodies toes or impede any other develpment. So again, I hope this is the right place for this and I apolagies in advance if it is not. 

Thank you 
Mark 




_______________________________________________ 
pve-devel mailing list 
pve-devel at pve.proxmox.com 
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel 






_______________________________________________ 
pve-devel mailing list 
pve-devel at pve.proxmox.com 
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel 


More information about the pve-devel mailing list