Overview

Overview is the first screen you see in Guardian — a fixed, zero-setup, fleet-wide snapshot of every registered device, refreshed in real time. It’s separate from Dashboards, where you build your own arrangement of widgets.

Connection status

The Live badge beside the page title, and the connection badge in the header, show how the page is receiving updates:

Badge	Meaning
Live	Connected to the real-time event stream
Connecting	Establishing the stream
Polling	Stream unavailable — falling back to a 30-second refresh

The header badge also shows Reconnecting, Live paused (the stream dropped and Guardian is retrying) and Offline (this device has no network connection).

Even when live, the dashboard also refreshes automatically every 30 seconds as a safety net.

Needs attention

Directly under the title, Needs attention lists what an operator should look at first, most serious first. It is built only from live data:

Open incidents (anything not yet resolved), by severity
Unresolved alerts, by severity
Offline agents — the first three by name, then a single “N more agents offline” row so a wide outage doesn’t bury everything else
Actions waiting for approval (technicians and above)

Each row opens the screen that owns the problem — an offline agent opens Agents filtered to it, an incident opens Incidents, and so on. When there is nothing to act on, the panel shows All clear with a one-line summary.

A freshness note (“Updated 12s ago”) sits beside the heading. If the page can’t refresh (for example the connection to Guardian drops), a banner says so and shows how old the data on screen is. The panel will not say “All clear” in that situation — it tells you the current status is unknown until Guardian responds, and Retry reloads the page’s data immediately.

Fleet summary

Four cards summarize the whole fleet, and every one is clickable:

Total Agents — registered endpoints; opens Agents
Agent Health — agents by state: Offline (no heartbeat for more than 30 seconds, opens Agents filtered to offline), Alerting (online with an unresolved alert, opens Alerts), and Healthy (online with no unresolved alert, opens Agents filtered to online)
Open Incidents — incidents not yet resolved; opens Incidents
Active Alerts — unresolved alerts across the fleet; opens Alerts

Below that, three cards cover:

Fleet Uptime — average uptime percentage over the last 30 days
Downtime This Month — total downtime across all agents
Downtime Cost — estimated financial impact, calculated from the hourly rate and currency configured in Settings

Resource usage and history

Gauges show current average CPU and memory usage across the fleet. Below them, CPU and Memory history charts plot fleet-wide usage over a selectable range: 1h, 6h, 24h, or 7d.

Registered Agents and Recent Alerts

A table lists every agent with its online/offline status, hostname, IP address, OS/architecture, and time since last seen. See Agents for the full device inventory. Alongside it, a panel lists the five most recent unresolved alerts by severity, title, and hostname — see Alerts for the complete list.

Historical Replay

Once you have at least one agent, a Historical Replay button reveals a slider that rewinds the entire dashboard — summary, charts, agent list, and alerts — to any point in the last 7 days. Dragging back near the present returns you to the live view automatically. Historical Replay requires an eligible plan.

Predictive Intelligence

If Guardian detects a device trending toward failure, a banner appears at the top of the dashboard highlighting the prediction so you can act before an outage occurs.

To build your own arrangement of widgets instead of this fixed view, see Dashboards.