Device health

The Health page#

Devices > Health is the whole fleet at a glance. The number beside Health in the sidebar is how many devices need a look. The page has two tabs: Overview, described here, and Battery health.

• Counts by health: Critical, Warning, Offline, Healthy and Not monitored. Click one to open the device list filtered to it. The bar under them also counts Asleep devices apart from Offline, because the agent said they went to sleep.
• Needs a look: critical and warning devices, and offline ones whose policy alerts on it, critical first. Each row shows CPU, memory, the fullest disk and what is wrong.
• Failing checks: services, websites, ports and event logs that are not passing, and for how long. Click one to open that check on its device.
• Disks nearly full: volumes at 80% or more, or forecast to fill within 30 days, with the forecast (for example Full in about 9 days, or Not filling).
• Monitoring alerts: the open alerts, critical first.

The page follows the customer switcher at the top of the screen, and the picker at the top right (Every device) narrows it to one device group.

Tip: Disk forecasts are a straight line through the last 14 days of hourly averages. Only rising trends get a date, and a weak trend says so.

Battery health#

The Battery health tab lists every laptop's battery against its design capacity, from the hourly battery health check, so you can plan which laptops to replace. The figures at the top count the laptops that reported, the average health, and how many are worn or due for replacement. Each row shows the laptop, its customer, the Last user, the model and the health. Filter by Customer, Last user and State.

A device's health#

On a device page's Overview tab, Health comes first. Use On this page to jump to any section below it.

• The badge shows the device's health, with chips for open alerts, failing checks, maintenance and mutes.
• CPU, memory, swap and each disk, with space free and how soon it fills when it is filling. What's using the space under a disk shows its biggest folders and files (see Script, file and server checks).
• USB disks, SD cards and network shares are listed apart under External and network. Mounted installer images and system volumes are never shown as disks. See Sleeping devices and volumes.
• Who is signed in, how long it has been up, when data last arrived and how often it is sampled. When the device is offline the numbers are dimmed and say how old they are.
• Charts for CPU, memory, each disk, the network (in and out) and, where reported, load and temperature. The device's own thresholds are drawn on them as dashed lines.

Using the charts#

• Pick a range: 1h, 6h, 24h, 7d, 30d, 90d or 1y.
• Drag across any chart to zoom in on that stretch. The range shows as a chip you can clear.
• Hover to see the exact time and value; the crosshair is shared across every chart on the page.
• Click a legend entry to hide or show a line (Option or Alt click shows it alone).
• The range goes into the page address, so you can send a colleague the exact view.

Gaps in a chart are real gaps in the data, such as a laptop switched off overnight.

Checks#

The Checks section lists every check that runs on the device, with its status, what it checks, the last result and the last 24 hours.

• Run checks now (or press C) runs every check at once; the play button on a row runs one.
• Add check adds a check for this device only, in its own override policy.
• Click a check to open its history.

A check that cannot apply to the device shows greyed out as N/A, with the reason beside it (for example, a hardware reading on a virtual machine). It never alerts and never counts against the device's health. A check that applies but could not run is a quiet notice until it has failed to run several times in a row. See Quiet monitoring and onboarding findings.

The history shows when it last ran and passed, how many times it has failed in a row, whether an alert is open, the pass and fail strip and value chart for the chosen range, and every result. Run now, Mute and Edit in its policy are at the top.

Hardware and security health on every system#

The agent reads the device's health on Windows, macOS and Linux where the thing exists: temperatures and fans, memory errors on machines with ECC memory, and security basics such as the TPM, Secure Boot and the screen lock (on Linux, the screen lock of the person signed in at the desktop). Where something does not exist, such as a TPM on a Mac or a screen lock on a server without a desktop, it says not applicable with the reason rather than failing.

Services#

The Services section lists every service the agent reports, those set to start automatically but not running first. Filter by All, Not running, Running or Stopped, by state and start type, or by name. A Linux service that waits for its socket (such as SSH started on demand) shows as Waiting for connections (socket), not Stopped.

Alerts and their history#

The Open alerts section lists the device's open alerts from every module, oldest problem first by when it opened. Click one to open it, with what raised it and any fix on offer (see What raised an alert, and fixing it from the alert). Switch to History to see every alert in a period, with the noisiest problem and the average time to acknowledge and resolve. See Alert history, dependencies and flapping.

Under the alerts, the device says what it Depends on (its firewall, switch or host), with Change.

The Monitoring card#

The Monitoring card in the side column lists the policies that reach the device (most specific last), its device groups, its maintenance windows and its mutes. It also has When it goes offline: Automatic (by the device's type), Always on (alert when it goes offline, for a kiosk or a reception PC) or Never alert, with a line saying why the device does or does not alert.

Click Config to see exactly how the device is monitored.

• With the agent confirms the agent has the current configuration.
• Policies lists them in order.
• Settings, Checks and Thresholds show each value and the policy it comes from.

From here you can change an inherited check or threshold for this device only (Tenvara makes a copy with the same key in the device's own policy), switch one off, go back to the inherited one, or add new ones with Add check and Add threshold.

The side column also shows Availability: the share of the last 30 days the device was reporting, with a strip by day.