-
v0.10.11 — remote machine cards showed the viewing node's process uptime
Release / version (push) Canceled after 0sRelease / frontend (push) Canceled after 0sRelease / linux (amd64) (push) Canceled after 0sRelease / linux (arm64) (push) Canceled after 0sRelease / windows (push) Canceled after 0sRelease / macos (push) Canceled after 0sRelease / release (push) Canceled after 0sBuild and Push Docker Image / build-and-push (push) Canceled after 0sLastSkywalker released this
2026-09-09 11:44:27 +03:00 | 63 commits to main since this releaseEvery remote host card (on a bridged hub and inside a cluster) showed one identical
uptime because a 10-minute backfill replicated host rows with boot_time=0 and /health
then fell back to this server's own process uptime.- Backfill carries each row's boot_time.
- A writer that doesn't know the boot time (0) never erases a known value; a ±2 s
now − uptime wobble keeps the stored value, a real reboot still lands. - Proxmox/PBS pollers stabilise the value before fingerprinting, so the wobble no
longer costs a Raft round per poll. - /health?host_id= reports no uptime when the boot time is unknown instead of the
process uptime.
Downloads