AtlasPVE: a safety layer for Proxmox that shows what an action touches (live demo, feedback welcome)

G

gokselurhan

New Member
Jul 13, 2026
3
0
1
Hi all,

I built AtlasPVE, a web control panel for Proxmox VE built around one idea: show you what an action touches before you run it. Full disclosure: I'm the developer and it's a commercial project, so this is a show-and-tell, not a sales pitch. I'd really like feedback from people who run Proxmox daily.

What it does differently:
- Before an update or a destructive action, it previews the impact (which VMs, storage, services are affected) instead of you finding out afterward.
- Guarded host updates: dry-run, an automatic snapshot before, and a boot-guard so a bad kernel can't strand you.
- Storage and topology maps (ZFS/LVM/disks, VM to bridge to NIC) so the layout is visible at a glance.
- A built-in watchdog for daily health and alerts.

Live demo, no install, click around: https://atlaspve.com

I'm not trying to replace the built-in UI for everyone. It's for the person who wants a deeper, safety-first daily cockpit. If you try the demo, I'd love to hear what's missing or what you'd never use. Honest criticism especially welcome.

Thanks.
 
Shouldn't this be clear by itself? I mean I can switch to nodes view in the ProxmoxVE WebUI and then I will see which node has which vms/lxcs and storage on it.

And why do you need a commercial product for snapshots if stuff like sanoid or zfs-auto-snapshot are available for free from Debian packages?
 
Johannes S said:
Shouldn't this be clear by itself? I mean I can switch to nodes view in the ProxmoxVE WebUI and then I will see which node has which vms/lxcs and storage on it.

And why do you need a commercial product for snapshots if stuff like sanoid or zfs-auto-snapshot are available for free from Debian packages?
Click to expand...
Thanks both are fair questions..

On the node view,. you're right, the built_in UI already shows what lives where. But what I'm after is a bit different. I'm not trying to show the current state, I'm trying to show what an action will touch before you run it. Before a host update or a storage change, Atlas spells out the impact: which guests are attached to this pool, whether a reboot is needed, what restarts. The node view shows you state, but it won't warn you about the operation itself. Whether that's useful to someone is fair to debate, not everyone needs the same thing,

On snapshots, I agree, sanoid and zfs-auto-snapshot are great, I'm not trying to replace them, use them by all means. Atlas isn't a snapshot scheduler. The pre-snapshot is just a seatbelt around risky operations like updates: dry_run, snapshot and boot-guard wrapped around that one action. What I'm really after is this: a beginner can grasp it easily, but an expert still finds enough depth. The reasoning is that even if you're an expert, on the deeper technical steps you want to cut the human factor down as much as you can, and get the job done fast and save time. Why not_?
 
LnxBil said:
Looks like a homelab project for homelabbers. How does it look and feel if you load your 32-node cluster with 1000+ VMs?
Click to expand...
this started for people running a single host or a small cluster who want to see their day to day work from a tidier place.. Not for an operation with 32 nodes and 1000+ VMs.. I haven't tested it at that scale and I won't claim it fits there. For big multi-cluster fleets like that, PDM or heavier platforms are probably the better call. I'd rather say it as it is than oversell and make promises I can't keep..

That said., federation is something we're thinking about, and honestly something we need to do.. With feedback from experienced people like you, the architecture and the flows could evolve in other directions.. I don't have it today, but why not..

Thanks for the criticism, this is exactly the kind of feedback I was hoping for
 
You must log in or register to reply here.
Top Bottom
Back