Introduction
You just finished installing Proxmox VE, rebooted the box, and typed the address into your browser: https://your-ip:8006. Now you're staring at a screen with a tree on the left, a dozen tabs across the top, and a log panel at the bottom that's already scrolling with tasks you didn't run. It's a lot at once.
Every Proxmox tutorial assumes you already know where things live. "Just go to the Firewall tab" doesn't help much if you've never actually found the Firewall tab. This one is different. We're not building a VM or fixing an error today — we're just walking around the interface so the rest of your Proxmox VE journey makes sense.
Once you know what each part of the screen does, every other Proxmox guide gets about 50% easier to follow.
What You Will Learn
- How to log into the Proxmox VE web interface and what that certificate warning actually means
- What the resource tree on the left is showing you, and the four ways you can view it
- What the Datacenter level controls versus what a Node controls
- The tabs you'll see on a virtual machine or LXC container, and what each one is for
- The difference between noVNC, SPICE, and xterm.js when you open a console
- A few CLI commands that mirror what the GUI is doing behind the scenes
- Common first-login errors and how to get past them
What Is This Feature?
The Proxmox VE web interface is the browser-based management console built into every Proxmox VE install. There's no separate app to download — it's served directly by the host on port 8006, and it's the same interface whether you're running a single homelab box or a fifty-node cluster.
Two words you'll see constantly: node and cluster. A node is just one physical (or virtual) Proxmox VE server. A cluster is a group of nodes joined together so they can be managed from one screen and, if you want, migrate VMs between each other. If you've only got one machine, you still have a "cluster" — it just has one member.
Underneath the GUI, everything is actually an API call. The web interface is a front end for the Proxmox VE REST API, which is worth knowing because it means anything you can click, you can eventually script.
Why Would You Use It?
You could technically manage Proxmox VE entirely from the command line over SSH. Some experienced admins do exactly that for scripted or repetitive tasks. But for day-to-day work — creating a VM, checking how much RAM a container is using, restoring last night's backup — the GUI is faster and a lot harder to mess up.
It also gives you things the command line doesn't show you at a glance: live graphs of CPU and memory, a task log so you can see if that backup job actually finished, and a console window so you can see your VM's screen without installing a separate VNC client.
Honestly, even people who are comfortable with Linux tend to live in this GUI for most tasks and only drop to a shell when something needs a command the interface doesn't expose yet.
Prerequisites
- A working Proxmox VE installation (8.x or 9.x — this guide covers the current 9.2 interface, but 8.x looks almost identical)
- Network access to the host on port 8006 from your computer
- The root password you set during installation, or another account with access
- Any modern browser — Firefox, Chrome, and Edge are all officially supported
If you haven't installed Proxmox VE yet, do that first. This guide assumes you're already looking at a login screen.
Step-by-Step Tutorial
Logging In for the First Time
Open https://your-ip:8006 in a browser. You'll almost certainly get a certificate warning first — click through it (in Firefox, "Advanced" then "Accept the Risk and Continue"). That warning isn't a sign anything is wrong. Proxmox VE generates its own self-signed certificate during installation, and browsers don't trust self-signed certs by default. You can replace it with a real one later using the ACME tab, which we'll get to.
On the login screen you'll see three fields: User name, Password, and Realm. Realm trips people up constantly. Leave it on Linux PAM standard authentication and log in as root with the password you set during install. There's also a "Proxmox VE authentication server" realm — that's for separate Proxmox-only user accounts you create later, not for root.
The Header Bar
Across the top you'll find the Proxmox logo and version number on the left, a search box in the middle, and four items on the right: a link to the official documentation, buttons to Create VM and Create CT, and your user menu.
The search box is more useful than it looks once your setup grows — type a VM name or ID and it jumps straight there instead of you scrolling the tree. The user menu (click your username, top right) is where you'll find TFA setup, password changes, your preferred language, and — my personal favorite — the theme switcher if the default light theme isn't your thing.
The Resource Tree
The panel on the left is the resource tree, and it's how you navigate between everything on your Proxmox VE setup. By default it's in Server View, which groups VMs and containers under the physical node they're running on. There are three other views, and switching between them is just a dropdown at the top of the tree.
| View | What it shows | When it's useful |
|---|---|---|
| Server View | Everything grouped by node | Default — good for small setups |
| Folder View | Grouped by object type (all VMs together, all storage together, etc.) | Comparing similar objects across nodes |
| Pool View | Grouped by pools you've created | Multi-tenant setups, separating clients or projects |
| Tag View | Grouped by tags applied to VMs/CTs | Ad-hoc grouping — "prod", "test", "backup-daily", whatever you like |
For a single homelab node, you'll rarely leave Server View. Once you're running a dozen VMs across a few nodes, Tag View starts earning its keep.
Datacenter: The Cluster-Wide Settings
Click Datacenter at the top of the tree and you're looking at settings that apply to the whole cluster, not just one machine. The tabs across the top change to: Search, Summary, Cluster, Options, Storage, Backup, Replication, Permissions, HA, ACME, Firewall, Metric Server, Notifications, and Support.
You won't touch most of these on day one, but it's worth knowing what they're for:
| Tab | What it's for |
|---|---|
| Summary | Cluster-wide health and resource graphs |
| Cluster | Join or create a cluster, see cluster status |
| Options | Cluster-wide defaults — keyboard layout, console viewer, migration settings |
| Storage | Storage pools available to the whole cluster |
| Backup | Scheduled backup jobs |
| Permissions | Users, groups, roles, and access control (RBAC) |
| HA | High Availability groups and managed resources |
| ACME | Free Let's Encrypt certificates, so you can drop that self-signed cert warning |
| Firewall | Datacenter-wide firewall rules and security groups |
| Notifications | Where alerts get sent — email, Gotify, webhooks |
The Node Level
Click on an actual node name in the tree (something like pve or whatever hostname you chose during install) and the tabs change again: Search, Summary, Notes, Shell, System, Updates, Firewall, Disks, Ceph, Replication, Task History, and Subscription.
Shell is the one I'd point out first — it opens a full terminal to that node right inside your browser, no SSH client needed. Notes is an underrated one too: it's a plain markdown box tied to that node where you can jot down things like "this node has the 10Gbe NIC" or "don't reboot before 6pm." Updates is where you refresh the package list and apply upgrades from the GUI instead of SSHing in to run apt update. Ceph only shows up once you've actually installed Ceph on that node — if you haven't touched Ceph, ignore it.
Virtual Machine and Container Tabs
Click on an actual VM or LXC container in the tree and you get a third set of tabs. VMs and containers overlap on most of these but differ on a couple:
| Tab | VM | LXC Container |
|---|---|---|
| Summary | Yes | Yes |
| Console | Yes | Yes |
| Hardware / Resources | Hardware (disks, CPU, RAM, NICs) | Resources (CPU, RAM, mount points) |
| Network | Configured under Hardware | Separate Network tab |
| Options | Yes | Yes |
| Backup | Yes | Yes |
| Snapshots | Yes | Yes |
| Firewall | Yes | Yes |
| Permissions | Yes | Yes |
A quick refresher since you'll see both terms everywhere: a VM (virtual machine) runs its own full kernel and operating system, fully isolated from the host. An LXC container shares the host's Linux kernel and is much lighter on resources, but it can only run Linux, not Windows. If you're not sure which one you need for a given task, that's usually the deciding factor.
Opening a Console
Click Console on any VM or container and Proxmox VE gives you a few ways to actually see the screen:
| Console type | What it is | Best for |
|---|---|---|
| noVNC | Browser-based VNC viewer, no install needed | Default choice, works everywhere |
| SPICE | A separate protocol with better performance and clipboard/USB support | Desktop VMs where you need smooth graphics or copy-paste |
| xterm.js | A plain terminal, not a graphical console | LXC containers and VMs with a serial console configured |
SPICE requires a small client app (virt-viewer) installed on your computer — it downloads a .vv file when you select it, which your SPICE client then opens. For most day-to-day work, noVNC is fine and needs zero setup.
The Task Log
Along the bottom of the screen is a panel that lists recent tasks — every backup, every VM start, every migration. Click any row to see the full output, which is often the fastest way to find out why something failed instead of digging through log files over SSH.
Commands Explained
The GUI is really just a nice wrapper around the Proxmox VE API and a handful of CLI tools. A few of these are worth knowing, especially for troubleshooting:
pveversion— prints the installed Proxmox VE version and package versions. Useful when a guide says "this only works on 9.x" and you need to check what you're actually running.pvesh get /cluster/resources— queries the same API the GUI uses, returning a JSON list of every node, VM, container, and storage in the cluster. Handy for scripting or just seeing raw data without clicking through tabs.systemctl status pveproxy— checks whether the service that actually serves the web interface is running. If the GUI won't load at all, this is the first thing to check.journalctl -u pveproxy -n 50— shows the last 50 log lines from that service, useful whenpveproxyis running but something's still broken.
Common Errors
A few things trip up almost everyone the first week:
- "Your connection is not private" / certificate warning. Expected behavior with the default self-signed cert. Not a bug.
- Login fails with correct password. Almost always the Realm dropdown — make sure it's set to "Linux PAM standard authentication" for the root account, not "Proxmox VE authentication server."
- GUI won't load at all, connection refused. Usually
pveproxyisn't running, or a firewall (on your network or the host itself) is blocking port 8006. - Console opens but shows a black screen. Common right after creating a VM before an OS is installed — that's normal. If it happens on a VM that used to work, check whether the display is set to a type your browser's noVNC can render (SPICE-only displays won't show in noVNC).
Troubleshooting
If the interface won't load, SSH into the node directly and run systemctl status pveproxy. If it's not active, start it with systemctl restart pveproxy and check journalctl -u pveproxy -n 50 for the reason it stopped.
If you can reach the login page but authentication keeps failing, double check for a locked Caps Lock key (it happens more than you'd think) and confirm the realm. You can also try logging in from a different browser or an incognito window — cached credentials or an old session cookie occasionally cause weird login loops.
If a specific VM's console won't connect while everything else works fine, check that VM's Hardware tab for its Display setting. Switching it back to the default (usually VirtIO-GPU or the standard VGA option) and restarting the VM fixes most of these.
For cluster setups where nodes show as greyed out or offline in the tree even though they're clearly running, check time sync between nodes with pvecm status — Corosync (the service that keeps cluster nodes talking to each other) is sensitive to clock drift, and even a few seconds of difference can cause a node to appear unreachable in the GUI.
Best Practices
Don't keep using root@pam for everything once you're past the initial setup. Create a separate user under Datacenter → Permissions with only the roles it actually needs — it's a smaller habit change than it sounds, and it means one compromised script or leaked password doesn't hand over the whole box.
Turn on TFA (two-factor authentication) for the root account from the user menu. It takes about two minutes with any TOTP app.
If you're running more than a handful of VMs, spend five minutes tagging them (production, test, whatever makes sense) so Tag View actually becomes useful instead of just another empty dropdown option.
Replace the self-signed certificate with a Let's Encrypt one through the ACME tab if your Proxmox VE host is reachable from the internet or you're tired of clicking through browser warnings. It's a one-time setup and it's free.
Skim the Notes field on each node once you've got more than one. Future-you, at 2am during an outage, will thank present-you for writing down which node has the flaky NIC.
Frequently Asked Questions
Why does Proxmox VE use port 8006 instead of the standard 443?
So it doesn't conflict with any other web service you might run on the same host. You can front it with a reverse proxy on 443 if you want a cleaner URL, but the GUI itself always listens on 8006.
Can I manage Proxmox VE from my phone?
Yes, the web interface is responsive and works in a mobile browser, though creating VMs and using the console is a lot more comfortable on a laptop or desktop.
Is there an official Proxmox VE mobile app?
No. Some third-party apps exist that use the API, but the mobile browser experience of the actual web GUI is generally the more reliable option.
What's the difference between the Datacenter Summary and a node's Summary?
Datacenter Summary aggregates health and resource data across every node in the cluster. A node's Summary only shows that one machine's CPU, memory, network, and storage.
Do I need to memorize CLI commands if I'm only using the GUI?
Not really, but knowing a handful — pveversion, systemctl status pveproxy, journalctl — saves a lot of time when the GUI itself is the thing that's broken.
Conclusion
None of this is complicated once you've clicked through it a few times — it's just unfamiliar the first time you log in. Now that you know the difference between the Datacenter tabs and the Node tabs, what the resource tree views are for, and which console option to reach for, the rest of the Proxmox VE tutorials out there will make a lot more sense. Go click around. It's genuinely hard to break anything just by looking.