Merge pull request #8 from JinwangMok/feat/scalex-dash

feat: scalex dash — multi-cluster Kubernetes TUI dashboard

Author: JinwangMok

CLAUDE.md

@@ -67,7 +67,7 @@ cargo fmt --check                        # Format check
 ## Key Patterns
 
 - **GitOps-First**: Post-bootstrap, ArgoCD manages all cluster state via ApplicationSets.
-- **Sync waves**: 0=ArgoCD/cluster-config, 1=Cilium/cert-manager/Kyverno/storage, 2=cilium-resources/cert-issuers/kyverno-policies, 3=tunnel/keycloak, 4=RBAC.
+- **Sync waves**: 0=ArgoCD/cluster-config, 1=Cilium/cert-manager/Kyverno/local-path-provisioner, 2=cilium-resources/cert-issuers/kyverno-policies, 3=tunnel/keycloak, 4=RBAC.
 - **Idempotent**: Every CLI operation safe to re-run.
 - **Pure Functions**: Rust CLI uses pure functions for HCL/inventory/vars generation. No side effects in generators.
 - **Secrets**: Created by CLI, stored in `credentials/` (gitignored). Templates in `credentials/*.example`.

README.md

@@ -68,7 +68,7 @@ spread.yaml (루트 Application)
 | 0 | ArgoCD, cluster-config |
 | 1 | Cilium, cert-manager, Kyverno, local-path-provisioner |
 | 2 | cilium-resources, cert-issuers, kyverno-policies |
-| 3 | cloudflared-tunnel, socks5-proxy, keycloak |
+| 3 | cloudflared-tunnel, keycloak |
 | 4 | rbac |
 
 ---
@@ -139,5 +139,5 @@ spread.yaml (root Application)
 | 0 | ArgoCD, cluster-config |
 | 1 | Cilium, cert-manager, Kyverno, local-path-provisioner |
 | 2 | cilium-resources, cert-issuers, kyverno-policies |
-| 3 | cloudflared-tunnel, socks5-proxy, keycloak |
+| 3 | cloudflared-tunnel, keycloak |
 | 4 | rbac |

docs/ARCHITECTURE.md

@@ -0,0 +1,70 @@
+# CLI Reference (`scalex`)
+
+The `scalex` CLI is the primary tool for managing the entire ScaleX-POD-mini platform — from hardware facts gathering through SDI virtualization to multi-cluster provisioning.
+
+## Build
+
+```bash
+cd scalex-cli && cargo build --release
+# Binary: target/release/scalex
+```
+
+## Core Commands
+
+| Command | Description |
+|---------|-------------|
+| `scalex facts --all` | Gather hardware info from all bare-metal nodes |
+| `scalex facts --host <name>` | Gather from a single node |
+| `scalex sdi init <sdi-specs.yaml>` | Virtualize bare-metal → resource pool → VM pools |
+| `scalex sdi clean --hard --yes-i-really-want-to` | Full infrastructure reset |
+| `scalex sdi sync` | Reconcile bare-metal changes (add/remove nodes) |
+| `scalex cluster init <k8s-clusters.yaml>` | Kubespray → multi-cluster provisioning |
+| `scalex secrets apply` | Generate and apply pre-bootstrap K8s secrets |
+| `scalex bootstrap` | Install ArgoCD + register clusters + apply spread.yaml |
+
+## Query Commands
+
+| Command | Description |
+|---------|-------------|
+| `scalex get baremetals` | Hardware facts table |
+| `scalex get sdi-pools` | VM pool status |
+| `scalex get clusters` | Cluster inventory |
+| `scalex get config-files` | Config file validation |
+| `scalex validate` | Pre-provisioning config validation (YAML parsing, IP/CIDR conflicts, pool mapping) |
+| `scalex status` | 5-layer platform status report |
+| `scalex kernel-tune` | Kernel parameter recommendations and diff |
+
+## Quick Reference
+
+```bash
+scalex facts --all                           # HW facts
+scalex sdi init config/sdi-specs.yaml        # VM pools
+scalex cluster init config/k8s-clusters.yaml # K8s clusters
+scalex secrets apply                         # Secrets
+scalex bootstrap                             # ArgoCD + GitOps
+scalex status                                # Full status
+scalex get clusters                          # Cluster list
+scalex sdi clean --hard --yes-i-really-want-to  # Full reset
+```
+
+## VM Resource Budget
+
+`scalex plan` computes VM specs from component budgets. Source: `resource_planner.rs`
+
+| Cluster Role | CPU (mc) | RAM (MB) | Disk (MB) | VM Spec (vCPU / GB / disk GB) |
+|--------------|----------|----------|-----------|-------------------------------|
+| **Management** (Tower) | 2,450 | 3,904 | 11,136 | **3 / 4 / 20** |
+| **Workload** (Sandbox) | 1,450 | 2,368 | 8,064 | **2 / 3 / 20** |
+
+> Included: base-os, k8s-cp, cilium, coredns, argocd(mgmt), cert-manager, kyverno, keycloak(mgmt), cloudflared(mgmt), local-path(workload)
+> VM conversion: `vCPU = ceil(mc/1000)`, `GB = ceil(MB/1024)`, `disk = max(ceil(MB/1024), 20)`
+
+## Config Files
+
+| File | Purpose |
+|------|---------|
+| `credentials/.baremetal-init.yaml` | SSH access to bare-metal nodes |
+| `credentials/.env` | SSH passwords/key paths |
+| `credentials/secrets.yaml` | Keycloak, ArgoCD, Cloudflare secrets |
+| `config/sdi-specs.yaml` | VM pool definitions (CPU, RAM, disk, GPU) |
+| `config/k8s-clusters.yaml` | Cluster definitions (mode, role, addons) |