v1.78.0

v1.78.0 (2026-10-30)

Release Highlights

• Promote per-drive filesystem capacity and SAS monitoring (#1060) - Closes #470, Closes #494, Fixes #1015, Fixes #1038, Closes #1054, Closes #1057, Closes #1058
  • Drives now report the free space of the filesystems they hold, SAS and SCSI drives gain several new monitoring surfaces, Unraid parity arrays are detected, and the device detail page says where its failure-rate figures come from.
  • Per-drive filesystem free space on the dashboard cards and the device detail page. A card shows a single Free figure when one filesystem sits on the drive alone, and a mount count otherwise; the detail page lists each filesystem with its own used, free and total bytes. Attribution walks the kernel holder tree, so partitions, LVM, md and LUKS all resolve, and it handles NVMe namespaces and Unraid array devices. Needs the filesystem collector enabled and the same COLLECTOR_HOST_ID on every collector.
  • SAS and SCSI monitoring: PHY error counters per port and PHY, the Verify Error Counter log page, and background scan results, lifetime temperatures and wear counters.
  • Unraid parity arrays detected from the key/value mdstat, reported as a single array.
  • The device detail view attributes Backblaze as the source of its drive failure-rate data, with the API's observed-thresholds field corrected and cross-referenced.
  • A non-advancing bucket paging offset, which would hang startup, is now asserted over the wire.