Every SSD carries two very different risks. One is a fuel gauge — write endurance — visible, slow, and honestly reported by every health tool. The other is a fuse: the controller, which reports nothing until the instant it blows. Checking health properly means reading the first and respecting the second.
The health percentage tracks flash wear beautifully — and says nothing about the controller and firmware faults that end most SSDs’ lives. Read the gauge; never mistake it for a guarantee.
Windows users: install CrystalDiskInfo and you’re done — every SATA and NVMe drive appears with its health percentage, temperature and full attribute table. Own-brand dashboards go a step further on their own drives: Samsung Magician, Crucial Storage Executive and WD’s Dashboard add lifetime-writes totals and firmware updates in plain English. Mac users get the one-word SMART verdict in Disk Utility, and the full story from smartctl (smartmontools, installable via Homebrew) — which reads Apple silicon’s internal storage too.
Whichever window you’re looking at, the readings that carry weight are the same four. Percentage Used — the official NVMe wear counter, counting up from 0 — or its friendlier inverse, the ‘health’ score. Total data written, judged against the TBW endurance figure on your drive’s spec sheet; most drives get replaced with the overwhelming bulk of it unspent. Available Spare — the pool of substitute blocks the controller retires worn flash into; it should sit at or near 100%, and a visible slide is the clearest wear warning the drive can give. Media & Data Integrity Errors — a counter that has one acceptable value, zero, regardless of what the headline percentage says.
Here’s the uncomfortable statistic from any recovery bench: the SSDs that arrive dead rarely arrive worn. They arrive with pristine wear figures and a controller that stopped answering — a firmware crash or chip-level electronics fault that takes the drive from perfect to absent between one boot and the next. The component that failed is the same component responsible for writing the health report, which is why no gauge flagged it. Treat this as a design property, not bad luck: an SSD’s health check tells you about wear, and only a backup protects you from the fuse.
The silver lining lives behind the failure. A blown controller usually leaves the NAND flash — and your files — physically intact, sealed inside a drive nothing will talk to. Firmware-level repair and direct chip reads exist for exactly this, reassembling data the controller can no longer serve: that’s the core of our SSD not detected work, and it’s why a vanished SSD is a recovery case, not an obituary.
Clean numbers: record today’s data-written figure and diary a re-check in six months — monitoring an SSD is genuinely that light-touch. Spare slipping or integrity errors above zero: the flash is telling the truth early for once; copy everything off while the drive reads willingly, replace it, and skip the temptation to ‘confirm’ with stress tests — each full-drive scan spends more of exactly what’s running out. Slowdowns, freezes, disappearing-and-returning: those are controller symptoms, and power-cycling roulette makes them worse; back up now, in one calm pass. Already vanished: stop reflashing, stop swapping cables in hope, and let a bench read it — our SSD & NVMe recovery service starts with a free diagnostic and a written quote. One SSD-specific footnote: TRIM erases deleted files’ blocks within moments on internal drives — so for accidental deletions on an SSD, the backup was always the plan.
No. Health reflects physical flash wear and the controller’s condition; a format just rewrites the file system on top and, if anything, spends a little more of the write budget. If a format appears to ‘fix’ an SSD that was misbehaving, the fault was logical — and worth taking as a warning shot rather than a cure.
Never — defragmentation is a hard-drive ritual that actively harms SSDs. There’s no moving head to optimise for, so it delivers zero benefit while burning through write endurance rearranging blocks. Windows knows this and quietly runs TRIM instead of a defrag on SSDs; leave it to it.
Not on a timer, but treat it as end-of-service. At exhausted endurance the drive typically shifts toward read-only behaviour and errors climb; anything still on it should come off immediately, while reads are good. What you shouldn’t do is keep it in service ‘until it actually dies’ — the final failure mode isn’t always graceful.
Controller, firmware and chip-level SSD recovery on the Bristol bench — free 48-hour diagnostic, written quote first, no fix, no fee on most jobs.