Definition
SMART (Self-Monitoring, Analysis and Reporting Technology) exposes health and diagnostic information from many hard drives and SSDs.
How it works
Available attributes depend on vendor and interface. They may report reallocated sectors, errors, operating time or estimated wear. NVMe health logs expose additional indicators including percentage used and media errors.
Real-world example
If a server reports I/O errors, reviewing SMART attributes and supported self-tests can help identify failing hardware. A global 'PASSED' status does not guarantee that a drive will not fail suddenly.
Limits and best practices
Interpret attributes according to vendor documentation, track trends and review RAID-controller alerts. Do not wait for a SMART failure when a drive is clearly malfunctioning, and always maintain independent backups.