HomeLab OS v1.0.16 (hotfix build 2)

Why a second v1.0.16?

This is the same app version (1.0.16, no version bump) — a hotfix release of the original v1.0.16 build. The v1.0.16 tag/name was already taken by the first release, so this second build ships under v1.0.16-hotfix.

What changed

Critical fix — backend crash loop when a Proxmox node is offline

PVE /nodes omits mem/cpu/disk/uptime for nodes that are offline. The provider assumed those fields always exist, which produced NaN metrics. better-sqlite3 stores NaN as SQL NULL, violating the NOT NULL constraints on the metrics table, so the backend crashed with:

SqliteError: NOT NULL constraint failed: metrics.ram_used_gb

and restarted forever.

• Guarded all numeric fields (ram/disk used + total, usage %) with a toFinite() helper — offline nodes now safely read as 0.
• Sanitized every numeric field in emitTick before persisting, so no NaN can ever reach SQLite again.
• Added a regression test that reproduces the exact production error without the fix and passes with it (backend test suite now 40/40).
• Removed the obsolete version: key from docker-compose.yml (Compose v2 warning).

Deploy

cd /opt/docker/HomeLab-OS   # or wherever you keep the repo
git pull
docker compose up --build -d

If your backend is already crash-looping, this will recover it on redeploy.

HomeLab OS v1.0.16

HomeLab OS v1.0.16 — real Proxmox provider + step-by-step SETUP.md

• ProxmoxMetricsProvider: MOCK_MODE=false + API token -> live discovery of nodes/VMs/containers, no agent.
• SETUP.md: step-by-step install (dashboard, Proxmox token, HTTPS, Telegram, SMTP, updating, troubleshooting).
• Contract/wiring: stats history, clusters, source name, last poll error; shared SQLite history; WS accepts any broadcaster; search reads live servers.
• New env vars: PROXMOX_HOST, PROXMOX_TOKEN_ID, PROXMOX_TOKEN_SECRET, PROXMOX_VERIFY_TLS, PROXMOX_POLL_INTERVAL_MS.
• Verified: 39/39 tests, typecheck, build, end-to-end smoke test incl. failure path.

HomeLab OS v1.0.15

HomeLab OS v1.0.15 — the initial release.

A premium, self-hosted Infrastructure Operating System — a live Network Operations Center (NOC) dashboard for your homelab.

Features

• Real-time telemetry over WebSocket (2s cadence) with animated UI
• Fleet grid, server detail charts, network topology map, alert feed
• Hardware sensor registry with live/Unavailable/Not-Available states
• Global search (⌘K) across servers, alerts, quick actions
• Settings: users, integrations, account (password/2FA/recovery), quick actions, access modes, feature flags, backups, audit log
• Popup modal system for all mutating settings actions
• Dark/Light appearance + six accent themes (Green default), persisted locally
• Semantic CSS design tokens; status colors (success/warning/critical) stay semantic in every theme
• Auth with roles/permissions, read-only/safe/emergency-lock modes, SMTP email, encrypted secrets
• dashboardctl CLI for admin password recovery and DB maintenance
• Docker Compose deployment + nginx SPA/API/WS proxy

Notes

• MOCK_MODE=true by default (simulated telemetry) — see README to wire real providers
• Source archives (zip/tar.gz) attached for this tag preserve the exact v1.0.15 code for rollback/reference