docs: adopt system-health-check workflow and host inventories

The workflow and the four host capability inventories move here from the home project. System maintenance is archsetup's domain, and this puts the thresholds TOML and both of its consumers in one repo.

The workflow's prose severity rules now cite the installed TOML keys, and the TOML wins on disagreement. Inventory paths point at docs/homelab-inventory/. A note routes routine ratio/velox checks to the maint CLI, leaving this workflow for the non-Arch hosts and forensic deep dives.

Only the #+HOSTNAME: capability inventories moved (ratio, velox, mybitch, truenas, plus the truenas hardware-specs asset they link). The personal gear records stay in home, and the home-side removal is handed off through its inbox.

1 files changed, 175 insertions, 0 deletions

diff --git a/docs/homelab-inventory/ratio-desktop.org b/docs/homelab-inventory/ratio-desktop.org
new file mode 100644
index 0000000..86dcd4d
--- /dev/null
+++ b/docs/homelab-inventory/ratio-desktop.org
@@ -0,0 +1,175 @@
+#+TITLE: ratio - Desktop Workstation
+#+DATE: 2026-01-27
+#+HOSTNAME: ratio
+
+* Automated Capabilities
+
+# Maintained by the system-health-check workflow. Manual edits may be
+# overwritten when the workflow detects drift between this drawer and the
+# live system. Machine-readable capability signals used to dispatch which
+# checks run on this host.
+
+:PROPERTIES:
+:FS:         btrfs
+:PM:         pacman
+:ORCH:       topgrade
+:SNAPSHOT:   snapper
+:MESH:       tailscale
+:BACKUP:     source
+:VIRT:       libvirt,docker,podman
+:INIT:       systemd
+:LAST_AUDIT: 2026-05-25
+:END:
+
+* Overview
+
+| Field       | Value                                    |
+|-------------+------------------------------------------|
+| Hostname    | ratio                                    |
+|-------------+------------------------------------------|
+| Type        | Desktop workstation                      |
+|-------------+------------------------------------------|
+| Maker       | Framework                                |
+|-------------+------------------------------------------|
+| Model       | Desktop (AMD Ryzen AI Max 300 Series)    |
+|-------------+------------------------------------------|
+| Board       | FRANMFCP06 v A6                          |
+|-------------+------------------------------------------|
+| Firmware    | INSYDE UEFI v03.03 (2025-09-16)          |
+|-------------+------------------------------------------|
+| OS          | Arch Linux                               |
+|-------------+------------------------------------------|
+| Kernel      | 6.12.67-1-lts                            |
+|-------------+------------------------------------------|
+| Desktop     | Hyprland (Wayland)                       |
+|-------------+------------------------------------------|
+| Previous ID | cogito (repurposed and renamed to ratio) |
+|-------------+------------------------------------------|
+
+* CPU
+
+| Field         | Value                                 |
+|---------------+---------------------------------------|
+| Model         | AMD Ryzen AI MAX+ 395 w/ Radeon 8060S |
+|---------------+---------------------------------------|
+| Architecture  | Zen 5 (TSMC 4nm)                      |
+|---------------+---------------------------------------|
+| Cores/Threads | 16 cores / 32 threads                 |
+|---------------+---------------------------------------|
+| Base/Boost    | 3.0 GHz / 5.15 GHz                    |
+|---------------+---------------------------------------|
+| Cache L1      | 1.2 MiB (d-16x48 KiB; i-16x32 KiB)    |
+|---------------+---------------------------------------|
+| Cache L2      | 16 MiB (16x1024 KiB)                  |
+|---------------+---------------------------------------|
+| Cache L3      | 64 MiB (2x32 MiB)                     |
+|---------------+---------------------------------------|
+| NPU           | XDNA 2, 50+ peak AI TOPS              |
+|---------------+---------------------------------------|
+| Peak Perf     | 59.4 FP16/BF16 TFLOPS @ 2.9GHz        |
+|---------------+---------------------------------------|
+
+* GPU
+
+| Field        | Value                                   |
+|--------------+-----------------------------------------|
+| Model        | AMD Radeon 8060S (Strix Halo)           |
+|--------------+-----------------------------------------|
+| Architecture | RDNA 3.5                                |
+|--------------+-----------------------------------------|
+| CUs          | 40 Compute Units                        |
+|--------------+-----------------------------------------|
+| Driver       | amdgpu (kernel)                         |
+|--------------+-----------------------------------------|
+| Max VRAM     | 96GB (via AMD Variable Graphics Memory) |
+|--------------+-----------------------------------------|
+| GPU Arch ID  | gfx1151                                 |
+|--------------+-----------------------------------------|
+| PCIe         | Gen 4, 16 GT/s, 16 lanes                |
+|--------------+-----------------------------------------|
+
+* Memory
+
+| Field        | Value                    |
+|--------------+--------------------------|
+| Total        | 128 GiB unified (LPDDR5) |
+|--------------+--------------------------|
+| Speed        | 8000 MT/s                |
+|--------------+--------------------------|
+| Channels     | 8 (8x 16 GiB)            |
+|--------------+--------------------------|
+| Manufacturer | Micron Technology        |
+|--------------+--------------------------|
+| Part Number  | MT62F4G32D8DV-023 WT     |
+|--------------+--------------------------|
+
+* Storage
+
+| Device       | Model                  | Size      | Interface |
+|--------------+------------------------+-----------+-----------|
+| /dev/nvme0n1 | WD BLACK SN850X 8000GB | 7.28 TiB  | NVMe      |
+|--------------+------------------------+-----------+-----------|
+| /dev/nvme1n1 | WD BLACK SN850X 8000GB | 7.28 TiB  | NVMe      |
+|--------------+------------------------+-----------+-----------|
+| *Total*      |                        | 14.55 TiB |           |
+|--------------+------------------------+-----------+-----------|
+
+Filesystem: btrfs with subvolumes (@, @home, @snapshots, @log, @pkg)
+
+* Network
+
+| Interface | Chipset         | Speed   | Type     |
+|-----------+-----------------+---------+----------|
+| enp191s0  | Realtek RTL8126 | 5 GbE   | Wired    |
+|-----------+-----------------+---------+----------|
+| wlp192s0  | MediaTek MT7925 | Wi-Fi 7 | Wireless |
+|-----------+-----------------+---------+----------|
+
+Also has: docker0, tailscale0, virbr0 (virtual interfaces)
+
+* Audio
+
+- AMD Radeon High Definition Audio (HDMI, snd_hda_intel)
+- AMD Ryzen HD Audio (3.5mm, snd_hda_intel)
+- PipeWire v1.4.10 (pipewire-pulse, wireplumber)
+
+* PCI Slots
+
+| Slot | Info       | Status    |
+|------+------------+-----------|
+| 1    | M.2 JWLAN  | Available |
+|------+------------+-----------|
+| 2    | M.2 JSSD1  | In use    |
+|------+------------+-----------|
+| 3    | M.2 JSSD2  | In use    |
+|------+------------+-----------|
+| 4    | PCIe Gen 4 | Available |
+|------+------------+-----------|
+
+* Connected Peripherals
+
+- Monitor: Dell U3419W (HDMI-A-1, 3440x1440@60Hz)
+- Keyboard: Das Keyboard 4 (Cherry MX Blue) — replacing with Keychron Q6 Pro
+- Mouse: Logitech M650
+- DAC/Amp: JDS Labs Element IV (ordered)
+
+* AI/LLM Capability
+
+Primary purpose includes local AI/LLM inference via Ollama.
+
+| Model            | Size  | Purpose              |
+|------------------+-------+----------------------|
+| deepseek-r1:70b  | 42 GB | Main reasoning model |
+|------------------+-------+----------------------|
+| qwen3:30b        | 18 GB | MoE chat (256K ctx)  |
+|------------------+-------+----------------------|
+| deepseek-r1:14b  | 9 GB  | Light reasoning      |
+|------------------+-------+----------------------|
+| nomic-embed-text | 274MB | RAG embeddings       |
+|------------------+-------+----------------------|
+
+Inference benchmarks (AMD testing, LM Studio 0.3.11):
+- Small models (1-3B): ~100+ tok/s
+- Medium models (7-8B): ~60-80 tok/s
+- Large models (20B): ~58 tok/s
+- Very large models (120B): ~38 tok/s