v0.21.9 — Disk Health & Life Prediction

Disk Health & Life Prediction

New DISK HEALTH section at the top of the IO page showing drive wear, age, and failure prediction.

Zero External Dependencies

• NVMe: Direct NVME_IOCTL_ADMIN_CMD kernel ioctl — reads the SMART/Health Information Log page directly from the NVMe controller firmware. No smartctl or nvme-cli needed.
• SATA SSD/HDD: Direct HDIO_DRIVE_CMD kernel ioctl — reads SMART attribute table directly. No smartmontools needed.
• Smartctl fallback: Uses smartctl --json for any devices not covered by direct ioctls.
• Virtual disk detection: Automatically skips QEMU, VMware, Xen, VirtualBox, Hyper-V virtual disks.

What It Shows

Field	NVMe	SATA SSD	SATA HDD
Life gauge bar	percentage_used	wear_leveling_count	n/a
Health verdict	critical_warning bitmap	realloc + pending sectors	realloc + pending
Temperature	NVMe health log	SMART attr 194	SMART attr 194
Total written	data_units_written	total_LBAs_written	total_LBAs_written
Write rate	TB/year from writes/hours	TB/year	TB/year
Est. remaining	wear rate extrapolation	wear rate extrapolation	power-on age
Available spare	NVMe-specific	n/a	n/a
Media errors	NVMe-specific	n/a	n/a

Life Prediction

Estimated remaining life is computed from actual hardware-reported data:

• wear_rate = percentage_used / power_on_hours
• remaining_hours = (100 - percentage_used) / wear_rate
• Color-coded: green (>1.5y), yellow (6mo-1.5y), red (<6mo)

Health Verdicts

OK → WARN (media errors, realloc sectors) → WORN (>70% used) → CRITICAL (>90% used, spare below threshold) → FAILING (NVMe reliability degraded) → FAIL (SMART health failed)

NVMe Critical Warning Flags

Decodes all 5 NVMe critical_warning bitmap flags:

• Available spare below threshold
• Temperature above critical composite threshold
• Reliability degraded (media/internal errors)
• Media placed in read-only mode (endurance exhausted)
• Volatile memory backup device failed