Dashboard: Overview, Home and the widget board

What used to be the dashboard is now two spaces on the desktop:

• Server → Overview (space 1 in Server mode): how the node is doing right now.
• Personal → Home (space 1 in Personal mode): your day at a glance.

Both show a set of hero blocks at the top and your widget board below them. Metrics are streamed live over a WebSocket and update without reloading the page.

Server Overview

• Greeting and status line. “Good morning, name” (afternoon, evening or night depending on the time) and a one-line verdict, either “All systems nominal.” or how many issues need attention, with the number of installed apps, the last verified backup and open alerts.
• Metric tiles. CPU, Memory, Disk and Net in, each with a live sparkline. Click a tile to open the Monitoring space.
• Molly insight (“Molly noticed …”). Appears when something needs your attention, such as a disk that will fill up soon, with the space Docker could free.
• Timeline. A live feed of what happens on the node: installs, sign-ins, scans, backups and alerts.

Personal Home

• Weather where you are. Click it to pick another city.
• Now playing: what’s playing here or on your other devices, with controls.
• Today: today’s calendar agenda, including recurring events.
• To do: a to-do list that syncs across your devices.
• Drive: how your Drive space is used (photos, docs, music…).
• Recent notes.

Customize

Each space has a Customize button (“Show on this space”) that turns the hero blocks and the widget board on or off:

Your choices are saved to your profile, so they follow you to every device.

The widget board

The board is a customisable 4-column grid: every widget can be resized anywhere from 1x1 to 4x4, reordered, or grouped into sections, which you can also reorder. Open the catalogue with Add widgets and click a widget or drag it onto the board. See Dashboard Customization for sizing and the right-click menu.

Widget catalogue

Widgets in bold are new in 0.11.0:

• CPU Cores: live per-core load bars.
• Molly Noticed: Molly’s proactive insight (disk forecast, reclaimable space).
• Timeline: the live event feed from Overview.
• Now Playing · Accretion: cover, equaliser and progress for what’s playing on this or another device.
• Today: today’s calendar agenda.
• Drive Usage: Drive space by Photos, Docs, Music and so on.

Per-User Layout Sync

As of 0.9.7, your dashboard layout is stored server-side and follows you across every device you log in from. There are two independent layouts, one for the Overview board and one for the Home board, and each syncs separately.

When you rearrange or resize widgets on one device, the change is saved to the server immediately. Any other device you open will pick up the new layout on the next load.

Offline cache: your browser keeps the last known layout in localStorage so the dashboard renders instantly while the server response is in flight. When the fresh layout arrives it replaces the cached copy.

Stale-version conflicts: Cloud OS uses optimistic concurrency for layout saves. If two devices write at the same moment — for example, you resize a widget on a phone and on a laptop in quick succession — the second write is rejected with a 409 Conflict response. The desktop shows a brief toast (“Dashboard updated elsewhere”) and reloads the server’s latest version so you are never left with diverged state.

The layout API is documented in the Dashboard Layout API section of the API reference.

Metrics

The Overview metric tiles and the System Health, CPU Cores and System Monitor widgets all read the same live feed.

How Metrics Are Collected

The Go backend polls system metrics every 10 seconds using the gopsutil library. Raw readings are pushed to connected frontends via WebSocket in real time. For historical charts, data is stored in SQLite at three resolutions:

• 1-minute intervals retained for the last 24 hours
• 1-hour intervals retained for the last 30 days
• 1-day intervals retained for the last year

A background rollup goroutine automatically compacts older data into the lower-resolution tiers.

Running Apps Widget

The Running Apps widget shows your installed apps as an icon grid with their state (running, stopped or error). Click an app to open its page. Everything installed is also listed in Launchpad under In orbit.

Disk Health

The disk health widget provides proactive storage monitoring by combining real-time usage data with predictive analysis.

SMART Data

When smartctl is available on the host, Cloud OS reads Self-Monitoring, Analysis, and Reporting Technology (SMART) attributes from your drives. The widget displays key indicators such as reallocated sector count, power-on hours, and temperature.

Disk Full Prediction

Cloud OS applies linear regression over the last 30 days of disk usage data to estimate when your storage will reach capacity. The prediction is displayed as “disk full in N days” on the dashboard.

Historical charts

Time ranges and long-term history live in the Monitoring space (Server mode, space 3). The range you choose there determines which resolution tier is read from the database.

Service Status Widget

The Service Status widget shows system services and their current state.

By default the widget filters to running services only so the list stays short. Click the All toggle in the widget header to see every service regardless of state.

Pagination is built in: the footer shows ← N/M → where N is the current page and M is the total number of pages. Use the arrows to scroll through the full service list without the widget overflowing its grid cell.

Top Processes Widget

The Top Processes widget lists the highest-CPU and highest-RAM consumers on the host. As of 0.9.7, it paginates in the same way as Service Status (← N/M →) so you can browse the full process table without resizing the widget.

Storage Overview Widget

The Storage Overview widget enumerates real filesystems on the host. It recognises the following filesystem types: ext4, xfs, btrfs, zfs, f2fs, and vfat (typically the EFI or /boot partition). Pseudo-filesystems such as tmpfs, devtmpfs, and squashfs are excluded from the list.

When more than one disk is present, the widget paginates one disk per page. Use the ← → arrows in the widget footer to move between disks. Each page shows the mount point, filesystem type, total/used/free space, and a fill-level bar.

Network — Per Interface Widget

The Network — Per Interface widget gives you per-adapter throughput history without relying on the aggregate Network metric card.

Interface tabs: a tab strip across the top of the widget lists every non-loopback interface detected on the host (for example eth0, enp3s0, wlan0). Click a tab to switch the chart to that interface.

RX/TX area charts: two stacked area charts display inbound (RX) and outbound (TX) throughput using 24 hours of bucketed history pulled from GET /api/network/history.

Range picker: a compact pill group in the widget header lets you zoom the chart window:

Selecting a shorter range increases detail for recent activity; selecting 24 h gives the full-day trend at lower resolution.

Command Palette

Press Ctrl/⌘+K anywhere to open the command palette. It jumps to any space or tool, opens or installs apps, runs system actions (appearance, lock, sign out, check for updates) and searches notes, files, photos, music and bookmarks. Matching is fuzzy, so “next” finds “Nextcloud”. See The desktop.

Troubleshooting

Metrics show 0% or are not updating

Check the live load indicator in the menu bar. If it doesn’t move, verify that the WebSocket connection is active. Open your browser developer console and check for WebSocket errors. The backend pushes metrics every 10 seconds, so values should update shortly after page load.

Overview or Monitoring loads slowly

If charts take a long time to render, the SQLite metrics table may have grown large. Cloud OS automatically compacts data, but you can check the database size:

du -sh /var/lib/quazzar/data/quazzar.db

If the database exceeds 500 MB, check that the rollup job is running by reviewing the application logs.