Features & workflows
This walks through each area of the app in the order you'd actually touch them: connect something, look at it, act on it, get warned when it needs attention. Everything below is a page in the left sidebar of the running app; try the live demo alongside this if you want to click through as you read.
Connections
A connection is a PVE cluster, a standalone PVE node, or a Proxmox Backup Server remote. Add one from Connections with a host, port, and either an API token (recommended: scope it to what Ferrum actually needs) or a username/password. Everything else in the app is built by polling these on an interval and rolling the results up into one inventory.
Overview & Dashboard
Overview is the fleet at a glance: total nodes and guests, aggregate CPU/memory/storage, and active alerts across every connection. Dashboard is the same idea but yours to arrange: drag, resize, and add from 20+ widgets (per-connection health, top guests by resource use, storage headroom, recent tasks, alert feed, and more). Layout is saved per user, so a shared login doesn't force one layout on everyone.
Inventory, Cluster & Topology
Inventory lists every node, VM, and LXC across every connection in one sortable, filterable table, with bulk actions (start, stop, migrate) selectable across multiple guests at once, even across different connections. Cluster drills into one cluster's nodes and quorum state; Topology draws the whole fleet as a connected graph instead of a table, which is the faster read when you're trying to see how guests are distributed rather than look one up by name.
Open a node from Node detail for its own resource graphs, or a guest's console from Console: VMs get a noVNC framebuffer, LXCs get an xterm.js shell, both running entirely client-side against Proxmox's own console endpoints.
Bulk operations
The Bulk Operations page is for scripted-feeling changes across many guests at once (start/stop/restart, snapshot, or migrate a filtered selection), with a preview of exactly what will run before you confirm it, and the result logged like any other action.
Storage, Pools & Backups
Storage and Pools show pool usage and Ceph health per connection and rolled up fleet-wide. Backups tracks job status and replication across native PVE storage and any PBS remotes you've connected, one view instead of switching between a PVE datacenter's backup tab and a separate PBS UI.
High availability, Firewall & SDN
HA shows HA groups and resources with their current state and failover history. Firewall covers cluster-level and per-node rules: add, reorder, or disable a rule and it's applied through the Proxmox API immediately, same as editing it there directly. SDN zones, VNets, and subnets are managed the same way, scoped per connection.
Alerts & automation
The Alerts page is where threshold rules live (CPU, memory, disk, or guest-state conditions, per connection or fleet-wide), plus config drift detection (something changed outside of Ferrum) and a fleet health score that rolls all of it into one number. Guest lifecycle policies (e.g. clean up VMs stopped past a certain age) and capacity forecasting run from the same page. Turn on scheduled health-digest emails from Settings if you want a daily summary instead of watching the page. Terraform and Ansible inventory export live here too, for teams that want the current fleet state as code without hand-maintaining it.
Webhooks
Webhooks lets you register outbound URLs that fire on real-time events (an alert firing, a backup failing, a guest changing state) with the payload shape and a delivery log so you can confirm the other end actually received it.
AI Assistant
The AI Assistant is a chat interface that can call the same tool catalog Ferrum's MCP server exposes: look things up, and (if the admin has allowed it) take action, like restarting a guest or acknowledging an alert. It works with any OpenAI-compatible provider you configure, or the built-in local Needle 2 model with no setup at all; see Configuration.
Users, Profile & Settings
Users (admin-only) manages accounts and roles. Profile is where any user manages their own password, theme/look preferences, and API keys. Settings (admin-only) is the control panel for everything not in config.yaml: notification channels, SSO, security policy, AI provider list, and the REST API/MCP enable switches covered in API, MCP & audit log.
Look & feel
Twelve presets (Enterprise, Proxmox-native, Terminal, Glass Flight Deck, Midnight, Paper, Glassmorphism, Neumorphism, Brutalist, Solarized, High Contrast, and Aurora) live under Settings > Appearance (or your own Profile if the admin allows per-user choice), each with a light and dark variant and eight accent colors. Density (comfortable vs. compact row spacing) is a separate, orthogonal setting for anyone managing a very long guest list.
Tasks & Audit Log
Tasks shows in-flight and recent Proxmox operations (migrations, backups, snapshots) with their live progress. Audit Log (admin-only) records who did what and when for every mutating action, regardless of whether it came through the UI, the REST API, or MCP; see API, MCP & audit log.