That should be a sign how important this is for the integration. Just a couple of DIYlers is just not enough. If there would be a huge demand (from paying customers), it would already be implemented, yet there is none and I do get that (see my other comments).So many supporters over all those years and it is still not implemented.
Importance lies in the eye of the beholder. Let’s not forget - there’s already a working DIY solution, making this an exceptionally low-hanging fruit: a quick win for Proxmox that serves a vocal, skilled community that’s at least partly the vanguard of tomorrow’s paying customers.
Which solution? You know that this is a commercial product, you cannot just copy & paste stuff from other people to implement it. There is a well defined and decades old workflow on how code has to submitted to PVE including a transfer of copyright (or whatever legal term applies here). If the author does not comply, it's not a low hanging fruit for Proxmox.
Thanks for the copyright primer - brand new concept to me! Naturally, anyone taking the initiative would need to sort that out. But let’s wrap this up before it turns into a full-blown crash course in legal basics.
Nope it's about priorities and return of investment (that's my guess at least). I'm willing to bet my homelab hardware that Proxmox Server Solutions GmbH (the company behind PVE, PBS and PMG) generates most of it's income not from us homelabbers (because like myself they are mostly freeloaders) but from companys, government entities or ngos having a support subscription for their software. For ProxmoxVE they are most likely customers which used VMWare before and migrated to ProxmoxVE due to Broadcoms obscene license fees. Most of these customers are not interested in typical homelab requests like the integration of all kinds of metrics in the GUI. They are interested in things like a VSphere kind of managment interface (the reason Proxmox Datacenter Manager is developed) or continuing to use their (imho way to expensive) enterprise storage SABs without sacrificing features which "just worked" with VMware ( FC or ISCSI connected SANs can only use LVM/Thick without integration from the storage vendor, but before ProxmoxVE 9 it's wasn' possible to do snapshots on them. And ProxmoxVE 9 is still in beta) or Veeam ( like Support for S3 Cloud storage, which Veeam has support for, PBS 4 Beta has too now). Obviouvsly Proxmox managment thought it more worthwhile to develop stuff like the Datacenter Manager or PBS S3 Support than integrating metrics, most corporate environments won't need (because they already have a monitoring system in place).
I agree with most of your conclusions, but you're overlooking small businesses that are paying customers too. Many of them operate with a very lean infrastructure—just a few virtualized hosts, no expensive enterprise hardware, and minimal features (perhaps just RAID 1 or RAID 5, but no HA, etc.). They just use the Proxmox admin interface because there is just this one IT guy who runs all things. Typically, they run their infrastructure in not perfectly climate-controlled environments. One of my customers even runs it in a cupboard - during exceptionally hot summers this can become a problem. And no, they don't want to set up and maintain another box - not even a Raspberry Pi - just to enable themselves to check on the temperature levels of their CPU, SDDs, HDDs.Nope it's about priorities and return of investment (that's my guess at least).
...
And development needs funding because software developers want to pay their rent and food just as you and me
As stated above: the code already exists. The efforts to quality check the prerequisites and the code are minimal. The copyright needs to be sorted out and that's about it.
Although somehow this thread has turned a tangent, I'll just add that PVE being open source I believe that Homelabbers are definitely part "of the audience" as they are an integral part of the testing & development - as this forum constantly proves.
Yes, If would be payed for support in running their infrastructure I would definitively tell them that they should run a monitoring system for their VMs even if they just have a "few virtualized hosts" . In the end it makes the life of their sole IT guy/girl easier since he/she will get notifications in case of a full disc/high cpu load etc/failing IT service. And if their hardware doesn't allow running another VM or LXC they have different problems. Software costs can't be an argument since most standard monitoring software is open source anyhow. This is at least true for Icinga2 and Zabbix. To be fair up to now I never used Zabbix but I read, that it's easier to setup and maintain then Icinga2 but doesn't scale that well so it's propably a good fit for your usecase. On https://www.zabbix.com/download_agents there are download links for an appilance qcow file (so can directly be used to setup a VM on a PVE host) and several Linux/Windows agent installation.
That's true but imho it's at best delusional to think that virtualization software is aimed mainly at homelabbers as a target audience. At least from a business case point of view. Now of course this doesn't mean that homelabbers can't use virtualization (far from it I'm doing it myself
). My point was an answer that this kind of feature would be benefical for homelabs and that it's not rational not to include it. This is the point I tried to refute: One don't need to like it, but from a business point of view it absolutely makes sense to priorize other features higher if one missing feature isn't much of a problem for most paying customers (who run their own monitoring system)Let's agree to disagree. You are right about the advantages of notifications etc. Nevertheless, this thread was and is just about temperature readings. And this data would be easily accessible and displayable on every Proxmox server - without any need for a dedicated monitoring system - may it be physical or virtual.
Seriously? It's hard to see how the countless person-hours contributed by the Proxmox community wouldn't factor into its business case. "Delusional" is a strong word - perhaps a bit of self-reflection is in order as well.
I go with Johannes. And i´ll add another argument. If anything in this direction is implemented, it should be "correct".
Would using a dedicated monitoring system simplify the "correct" selection of relevant temperature sensors compared to doing so through the Proxmox interface? Please explain.
Monitoring systems have already all the interfaces for grouping, alarming, thresholds etc.; and all such special features, monitoring needs.
My question was purely rhetorical - with just a dash of sarcasm. Couldn't resist, sorry!
Oh, absolutely - those few lines are clearly wreaking havoc on the entire Proxmox UI. I mean, who needs a functional built-in monitoring section anyway when you're already preaching the gospel of dedicated monitoring tools? Consistency is overrated, right?
