Check an SSD’s health by identifying the exact drive, reading its reported warnings and wear indicators, and choosing an action based on both the report and the symptoms. If files are disappearing, reads fail or Windows reports a critical warning, protect the data before running tests. A green label or “99% health” is not a guarantee that every file is safe.
Start with the situation, not a benchmark
- Critical warning, read-only state or repeated read errors: stop avoidable writes and back up readable important files to separate storage. If irreplaceable files cannot be read reliably, stop repeated scans and consider professional recovery.
- Working normally, but the health percentage has fallen: save the current report and investigate the field behind that number. Wear is not a remaining-days clock.
- The SSD is missing or Windows asks to format it: use our drive-detection guide first. Formatting is not a health check, especially when the drive should contain data.
For a failing drive, do not start with CrystalDiskMark, Secure Erase, firmware flashing or a repair command that writes to the disk. Finding out how fast it can write matters less than preserving the only copy of your files.
Identify which SSD you are checking
Open Task Manager → Performance → Disk and record the model and capacity. With several drives, match the volume you care about to the correct physical disk; a report from the healthy second SSD does not clear the system drive.
Record the connection too: internal NVMe, internal SATA, or external USB. NVMe and SATA use different reporting paths. An enclosure or storage controller can prevent a utility from seeing details even while files remain accessible.
Keep the model, firmware version and tool version in your note. Redact the serial number before sharing screenshots publicly.
Check Windows, then a tool that supports the drive
In Windows 11, open Settings → System → Storage → Advanced storage settings → Disks & volumes. Select the physical SSD and open Properties. Look for drive-health information and any warning. Menu wording differs in older Windows versions.
What you should see: the selected disk’s identity and, on a supported setup, its health fields. Microsoft says this monitoring covers NVMe SSDs, not SATA SSDs or HDDs. A missing health section is therefore not a clean bill of health.
For more detail, use the manufacturer’s official utility for your model, or CrystalDiskInfo from Crystal Dew World. Open the drive’s status/SMART view and match its model before interpreting anything. CrystalDiskInfo reads condition information; CrystalDiskMark is a different program for performance benchmarking.
Samsung Magician varies its features by product; an OEM Samsung drive need not have the same support as a retail model. Kingston SSD Manager does not support USB enclosures. If a tool cannot read a field, record unavailable, not zero. Do not switch storage-controller modes merely to expose telemetry.
Read the fields that change your decision
Critical Warning and reliability messages
Start here. A spare-capacity, reliability or read-only warning needs attention even if another screen looks reassuring. Save the exact warning, secure the readable data and contact the drive or PC maker. If two tools disagree, confirm they are reading the same physical disk and use its manufacturer’s interpretation.
Percentage Used and remaining life
These labels point in opposite directions: used rises as estimated endurance is consumed; remaining falls. Neither is a calendar forecast. Microsoft’s NVMe field definitions explain that Percentage Used can reach or exceed 100 without proving immediate device failure.
That is a reason to plan replacement and maintain a verified backup, not permission to ignore warnings. Do not calculate days left by dividing a percentage by the SSD’s age.
Available Spare, errors and temperature
Available Spare describes reserve capacity inside the drive, not free space on C:. Compare it with the drive’s reported spare threshold. Media/data-integrity errors and their change over time deserve more attention than a percentage alone. An unsafe-shutdown counter records another kind of event; it is not interchangeable with media errors.
Record temperature with its context, such as idle or a normal workload. Use the exact model’s specifications and warning state rather than one “safe temperature” for every SSD and sensor.
SATA attributes and the overall label
Do not apply a random list of NVMe or HDD attributes to every SATA SSD. Wear labels, raw units and thresholds can be vendor-specific. The developer’s CrystalDiskInfo health documentation shows different SSD interpretations. A health color summarizes rules; it does not verify file contents or promise future reliability.
Which SSD signal changes the next action?
A wear percentage, a warning and an unreadable file answer different questions. Start with reported warnings and symptoms, not a color.
- Protect the only copy
Read failures or a critical warning: avoid unnecessary writes and secure readable files.
- Identify the field
Percentage Used is endurance consumption; Available Spare is internal reserve, not free space.
- Compare the same drive
Record model, firmware, tool and date. Track changes rather than guessing days left.
Teaching case A: Percentage Used 8, Available Spare 100, no reported warning. These fields alone do not establish an urgent failure, but also do not certify the files or replace a backup. Teaching case B: Percentage Used 8 with a reliability warning and unreadable files. The low wear estimate does not cancel the warning: protect data before further tests. Teaching case C: health unavailable through a USB enclosure. Record “unavailable”; do not turn it into 100% health or zero errors.
Runs only in your browser. Nothing is uploaded or saved. Use the written method above to collect the right readings.
NVMe Percentage Used can reach or exceed 100 without proving immediate failure. A reliability warning can matter at any wear percentage. Never estimate remaining days from one health number.
Keep a report you can compare
Copy this record into a note, or use Save as PDF to keep the guide beside the report:
Date / symptoms:
Exact SSD model / capacity / firmware:
Connection and status-tool version:
Backup location / sample files opened successfully:
Exact warning or overall status:
Wear field name / value / unit:
Available spare / threshold, if reported:
Media errors / temperature / workload:
Previous report compared / what changed:
Next action and when to check again:
Verify the backup by opening representative copied files. A second folder on the same physical SSD is not a separate backup. For a normally working drive, compare later reports made with the same tool and connection; repeated benchmarks add little to this question.
What to do next
Warnings or unreliable reads: prioritize data, then recovery/support and replacement planning. Wear without errors or symptoms: maintain backups, track the report and plan around the drive’s endurance and importance. No warnings and normal behavior: continue sensible backups; health monitoring cannot detect every future failure.
If slow performance is the only symptom, check free space, workload, cooling and connection limits separately. A health percentage does not identify the cause of every slowdown. Once replacement is justified, our SSD guide explains the compatibility questions to check.
Common questions
Is 98% health bad?
Not by itself. Find out which wear indicator produced it, whether warnings/errors exist and whether the drive behaves normally. Two different models’ percentages are not a precise reliability ranking.
Does “Good” mean I can delete my backup?
No. A status report is a snapshot of what the drive and software expose. It cannot protect against accidental deletion, theft, controller failure or every unreadable file.
Can software restore worn flash to 100%?
Changing an alert threshold or clearing a display does not restore physical endurance. Treat unexplained jumps as a reason to check the tool, firmware and model interpretation, not evidence that the SSD repaired itself.