Previously, the agent shared a single PrevReadTime timestamp across all
collection intervals (e.g., 1s and 60s). This caused the 60s collector
to divide its accumulated 60s byte delta by the tiny time elapsed since
the last 1s collection, resulting in astronomically inflated network
rates. The fix introduces per-cache-time read time tracking, ensuring
calculations for each interval use their own independent timing context.
- Match FILESYSTEM directly against I/O devices if partition lookup
fails
- Fall back to the most active I/O device if no root device is detected
- Add WARN logs in final fallback case to most active device
findIoDevice now normalizes device names and falls back to prefix-based
matching when partition names differ from IOCounter names (e.g. nda0p2 →
nda0 on FreeBSD). The most-active prefix-related device is selected,
avoiding the broad "most active of all" heuristic that caused Docker
misattribution in #1737.
The previous behavior only caught some errors including inaccessible
hosts, but not others like failed authentication or service
unavailability. This largely applies when using a socket proxy and
having the retry mitigates some erroneous behavior.
* SMART: add eMMC health via sysfs
Read eMMC wear/EOL indicators from /sys/class/block/mmcblk*/device and expose in SMART device list. Includes mocked sysfs tests and UI tweaks for unknown temps.
* small optimizations for emmc scan and parsing
* smart: keep smartctl optional only for Linux hosts with eMMC
* update smart alerts to handle warning state
* refactor: rename binPath to smartctlPath and replace hasSmartctl with smartctlPath checks
---------
Co-authored-by: henrygd <hank@henrygd.me>
- Stop using max-read fallback when mapping root filesystem to
diskstats.
- Keep root usage reporting even when root I/O cannot be mapped.
- Expand docker fallback mount detection to include /etc/resolv.conf and
/etc/hostname (in addition to /etc/hosts).
- Add clearer warnings when root block device detection is uncertain.
Changes container network statistics to use raw byte values instead of converting to megabytes agent-side, providing more accurate measurements for low-bandwidth containers. Maintains backward compatibility with older agents/hubs through fallback logic.
- Agent now sends Bandwidth field as [sent_bytes, recv_bytes] array
- Deprecated NetworkSent/NetworkRecv fields still populated for compatibility
- Hub and frontend fall back to deprecated fields when Bandwidth is zero
- Record averaging correctly handles both old and new formats
- TODO markers added for cleanup in version 0.19+
fixes#1130
* add RuntimeIdentifier and AppendRuntimeIdentifierToOutputPath to beszel_lhm.csproj
* add more default sensor filters for LHM
---------
Co-authored-by: henrygd <hank@henrygd.me>
- Move SELinux context handling to internal/ghupdate for reuse
- Make chcon a true fallback (only runs if semanage/restorecon unavailable)
- Handle existing semanage rules with -m (modify) after -a (add) fails
- Apply SELinux handling to both agent and hub updates
- Add tests with proper skip behavior for SELinux systems
---------
Co-authored-by: henrygd <hank@henrygd.me>
* feat: add ATA Device Statistics parsing and fall back for SMART temp reading
* simplify ata device statistics structs and fix smartctl args tests
* simplify ata device statistics lookup to use page number only
---------
Co-authored-by: henrygd <hank@henrygd.me>
