Backblaze’s most recent drive study covered around 443,000 drives across their fleet and pulled the same conclusion again: some drive families fail twice as often as their siblings on the same shelf. Consumer setups run one or two disks, not 443,000, but the math still applies. The difference between “I saw it coming” and “the drive is dead” is a couple of lines of SMART data your operating system already collects and never shows you.
We tested seven tools that read SMART, model failure risk, and yell before the disk gives up. The list spans a single-machine Windows box, a NAS, and a Proxmox rack, because the right app depends on where the drives live.
What matters in an HDD monitoring tool
- Reads SMART directly, not just OS-cached values. The OS often hides bad attributes.
- Supports HDDs, SSDs, and NVMe. Attribute meaning changes across drive types.
- Warns before failure, not at failure. Rising reallocated sectors are the tell.
- Long tests without babysitting. Extended self-tests take hours; the tool should schedule them.
- Notifies through more than a taskbar bubble. Email, webhook, ntfy, Home Assistant.
- Works with your enclosure. USB-to-SATA bridges hide SMART unless the tool speaks the bridge’s language.
Quick comparison
| App | Best for | Platforms | Free plan | Rating |
|---|---|---|---|---|
| CrystalDiskInfo | Single-Windows-machine daily driver | Windows | Free (donation ver.) | 4.8 |
| Scrutiny | NAS or homelab dashboard for many drives | Linux, Docker | Free (MIT) | 4.7 |
| smartmontools | Server, script, cron, headless | Windows, macOS, Linux, BSD | Free (GPL) | 4.7 |
| Hard Disk Sentinel | Full HDD test and repair suite on Windows | Windows, Linux | around $30 one-time | 4.6 |
| GSmartControl | Cross-platform GUI over smartctl | Windows, macOS, Linux | Free (GPL) | 4.5 |
| DiskCheckup | Passmark’s read-out with alerts | Windows | Free for personal | 4.3 |
| HD Tune | Fast benchmarks with light monitoring | Windows | Free basic, Pro paid | 4.4 |
The apps
1. CrystalDiskInfo — Best for a Windows daily driver
CrystalDiskInfo is the app that made SMART readable. Green-yellow-red status, plain-language attribute descriptions, health percentage, and temperature history. The tray icon changes color when an attribute crosses a threshold, and the notification path covers Windows Toast and email. Free, actively maintained, and the same tool the enthusiast community has installed since 2008.
Where it falls short: Windows only. The “Shizuku Edition” installer bundles anime themes; grab the Standard build if that is not your speed.
Pricing:
- Free: Everything
- Paid: None (author accepts donations)
Platforms: Windows
Download: crystalmark.info · GitHub mirror
Bottom line: Install this on every Windows PC with a drive you care about.
2. Scrutiny — Best homelab and NAS dashboard
Scrutiny sits between smartd and a browser. Deploy the Docker container, point it at every drive, and get a web dashboard with per-drive attribute graphs, temperature trends, and Backblaze-style failure prediction based on the same attributes their fleet study identified as leading indicators. Ntfy, webhook, and email notifications built in.
Where it falls short: Docker deployment is the only well-supported path. Bare-metal installs work but need manual smartd setup.
Pricing:
- Free: MIT-licensed
- Paid: None
Platforms: Linux (Docker on any host)
Download: github.com/AnalogJ/scrutiny
Bottom line: The pick for a NAS or Proxmox rack with more than two drives.
3. smartmontools — Best headless and scriptable
smartmontools is the CLI everyone else wraps. smartctl reads any attribute; smartd runs as a service and mails on threshold. If you already run cron or systemd timers, this is the tool that fits without ceremony. Every Linux distribution ships it. It also has clean Windows and macOS builds.
Where it falls short: No GUI. Attribute output is dense.
Pricing:
- Free: GPL-licensed
- Paid: None
Platforms: Windows, macOS, Linux, BSD, Solaris
Download: smartmontools.org · GitHub
Bottom line: The default for any server. Everything else is a GUI on top.
4. Hard Disk Sentinel — Best for a repair-focused workflow
Hard Disk Sentinel goes past monitoring into full disk testing and surface repair. Extended tests, reallocated-sector diagnostics, and destructive sector-remap tools live in the same app. If a drive has already started to slip, this is the tool that squeezes another few weeks out of it while you wait on the RMA.
Where it falls short: Windows-focused, and the license is per-machine. The Linux build lags features.
Pricing:
- Free: 30-day trial
- Paid: Around $30 for a Standard license, one-time
Platforms: Windows, Linux
Download: hdsentinel.com
Bottom line: Worth the $30 when a drive is on the edge and you want to know exactly where.
5. GSmartControl — Best cross-platform GUI
GSmartControl is a proper Gtk GUI over smartctl. It shows the same attributes but with tables, health verdicts, and buttons for tests. On Linux it beats every other GUI for actually running short and extended tests without fighting a config file.
Where it falls short: Interface is dated. No notifications.
Pricing:
- Free: GPL-licensed
- Paid: None
Platforms: Windows, macOS, Linux
Download: gsmartcontrol.sourceforge.io
Bottom line: The right pick on Linux desktops when the CLI is not your thing.
6. DiskCheckup — Best simple Windows monitor with alerts
DiskCheckup is Passmark’s take: a clean panel per drive, threshold alerts by email, and a temperature history. Lighter than CrystalDiskInfo, more approachable than smartctl.
Where it falls short: No support for non-Passmark USB enclosure quirks; some external drives read as unknown.
Pricing:
- Free: Personal use
- Paid: Site license for business use
Platforms: Windows
Download: passmark.com
Bottom line: Good pick when CrystalDiskInfo is too busy and smartctl is too dry.
7. HD Tune — Best when you want benchmarks in the same window
HD Tune started as a benchmark tool and grew into a monitor. Sequential read speed, access time, and error scan sit alongside SMART health. The Pro version adds file benchmarks and a secure erase.
Where it falls short: Development is slow. Newer NVMe drives sometimes show as unsupported until an update.
Pricing:
- Free: HD Tune 2.55 basic
- Paid: HD Tune Pro around $35 one-time
Platforms: Windows
Download: hdtune.com
Bottom line: The pick when you also want to see if the drive is slow, not just failing.
How to pick the right one
- If you run one or two Windows PCs: CrystalDiskInfo.
- If you run a NAS, homelab, or Proxmox cluster: Scrutiny.
- If you already script your infrastructure: smartmontools with smartd, and skip everything else.
- If a drive is showing errors and you want to fight for it: Hard Disk Sentinel.
- If you are on Linux and want a GUI: GSmartControl.
- If you want simple thresholds and email alerts on Windows: DiskCheckup.
- If you also care about benchmark performance: HD Tune.
FAQ
What does SMART actually predict? Attributes like reallocated sector count, current pending sectors, and uncorrectable errors correlate strongly with failure within weeks or months. Head fly-height and read error rate correlate more weakly. Backblaze’s public analysis identifies five attributes worth watching hardest.
Do these tools work on NVMe SSDs? Yes. All seven read NVMe SMART attributes on modern hardware. The attribute names are different from spinning drives and the tool has to translate; CrystalDiskInfo and smartmontools are best at this.
Which of these works over USB enclosures? smartmontools has the widest SAT translation database, so USB drives that hide SMART tend to work best there. CrystalDiskInfo supports many popular bridges through its config file.
Do I still need a monitoring tool if I have RAID? Yes. RAID protects against a drive dying; it does not prevent the drive from dying. Catching the drive on the way down saves a resilver.
How often should I run an extended self-test? Once a month on spinning disks that hold anything you care about. NVMe self-tests are shorter and can run weekly with no measurable impact.