Features · Hardware health
Hardware health
The early warnings your dedicated servers already give
Physical servers report failing disks, degraded arrays, memory errors and power faults long before they fail outright — to sensors nobody is reading. The Astrari agent reads them every 15 minutes, changes nothing, and raises a finding the moment something needs attention.
What it does
Read-only, every 15 minutes
The agent reads the server's own sensors and controllers through standard tools such as ipmitool, smartmontools and Dell's PERC utilities. It never runs disk self-tests or changes configuration.
Redundancy, not just health
A healthy single disk is still a single point of failure. Astrari checks whether the root filesystem and your main data volume would survive one disk failing, and tells you when they wouldn't.
Alerts that match the risk
Uncorrectable memory errors, a failed power supply or a degraded array alert as critical; rising corrected-error counts or a disk wearing out raise earlier, lower-severity warnings so you can plan the replacement.
Separate from the security score
Hardware findings are tracked and alerted like any other finding, but they don't affect the security score — a worn disk isn't a vulnerability, and the score stays meaningful.
What it watches
- ▸ Power supply failures and lost input power
- ▸ Failed or failing fans, and temperatures past their limits
- ▸ Corrected and uncorrectable memory errors, per DIMM where the hardware reports it
- ▸ Degraded, rebuilding or offline RAID arrays, and failed or predicted-to-fail disks
- ▸ SMART health, reallocated and pending sectors, NVMe wear and media errors
- ▸ Volumes with no disk redundancy
Example finding
RAID virtual disk degraded — Disk.Virtual.0
The array is running on one disk after its mirror reported a failure. Data is intact but no longer protected; replace the failed disk and the controller will rebuild.
How it works
- 1
The agent detects physical hardware
On cloud and VPS servers the provider monitors the hardware, so Astrari detects the virtual machine and skips these checks.
- 2
Add the standard tools if needed
If ipmitool or smartmontools is missing, install it from the server's Hardware section. Nothing is installed without your click.
- 3
Sensors are read every 15 minutes
Changes are reported as they happen, with an hourly heartbeat. A sensor that can't be read is shown as unreadable — it never clears a finding.
- 4
Faults become findings and alerts
Each issue opens a Hardware finding on the server and alerts through your usual channels, and resolves itself when the component reports healthy again.
Plans
Hardware health is included on every plan, for physical servers. Compare plans.
Good to know
- ▸ If a server loses power or its board fails, the agent stops too and you'll receive an agent-offline alert. For warnings that survive a dead server, enable email alerts in the server's iDRAC, iLO or BMC.
- ▸ Dell PERC arrays are read with perccli, storcli or racadm when installed; these are downloaded from Dell's support site.
See it on your own servers
The free Starter plan covers one website and one Linux server. The agent installs with one command and reports its first scan within a minute.