chore(infra): remove llm-api service (retired from prod)

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
fix(hamc): hermyhq port 4023→4024, add hermyhq-render worker unit
2026-09-25 18:53:29 +07:00 · 2026-09-22 20:16:33 +07:00 · 2026-09-21 13:38:04 +07:00 · 2026-09-20 21:48:22 +07:00 · 2026-09-20 15:51:34 +07:00 · 2026-09-02 12:31:18 +07:00
7 changed files with 110 additions and 25 deletions
@@ -1,4 +1,4 @@
-name: Deploy Caddy Config
+name: Deploy Infra Config

 on:
  push:
@@ -7,10 +7,11 @@ on:
      - 'infra/caddy/**'
      - 'infra/firewall/**'
      - 'infra/systemd/**'
+      - 'infra/prometheus/**'
  workflow_dispatch:

 concurrency:
-  group: caddy-deploy
+  group: infra-deploy
  cancel-in-progress: false

 permissions:
@@ -21,16 +22,15 @@ env:
  VPS_USER: ${{ secrets.VPS_USER }}

 jobs:
-  deploy-caddy:
+  deploy-infra:
    runs-on: ubuntu-latest
-    timeout-minutes: 10
+    timeout-minutes: 15
    steps:
      - uses: actions/checkout@v7

      - name: Validate Caddyfile syntax
        run: |
-          # Real Caddy parser (better than naive brace counting)
-          curl -fsSL https://caddyserver.com/api/download?os=linux\&arch=amd64 -o /tmp/caddy
+          curl -fsSL "https://caddyserver.com/api/download?os=linux&arch=amd64" -o /tmp/caddy
          chmod +x /tmp/caddy
          /tmp/caddy validate --config infra/caddy/Caddyfile.prod --adapter caddyfile 2>&1 | tail -5
          echo "✅ Caddyfile valid"
@@ -49,12 +49,39 @@ jobs:
      - name: Sync Caddyfile to VPS
        run: |
          set -e
-          # Backup current config, then push the new one
          ssh "$VPS_USER@$VPS_HOST" "sudo cp /etc/caddy/Caddyfile /etc/caddy/Caddyfile.bak-previous"
          scp -q infra/caddy/Caddyfile.prod "$VPS_USER@$VPS_HOST":/tmp/Caddyfile.new
          ssh "$VPS_USER@$VPS_HOST" "sudo cp /tmp/Caddyfile.new /etc/caddy/Caddyfile && sudo rm -f /tmp/Caddyfile.new"
          echo "✅ Caddyfile synced"

+      - name: Sync systemd drop-ins to VPS
+        run: |
+          set -e
+          if [ -d infra/systemd ]; then
+            for f in infra/systemd/*; do
+              [ -f "$f" ] || continue
+              base=$(basename "$f")
+              echo "  syncing $base"
+              scp -q "$f" "$VPS_USER@$VPS_HOST":/tmp/"$base"
+              ssh "$VPS_USER@$VPS_HOST" "sudo mkdir -p /etc/systemd/system && sudo cp /tmp/$base /etc/systemd/system/$base && sudo rm -f /tmp/$base"
+            done
+            ssh "$VPS_USER@$VPS_HOST" "sudo systemctl daemon-reload"
+            echo "✅ systemd drop-ins synced"
+          else
+            echo "no infra/systemd/ files"
+          fi
+
+      - name: Sync prometheus targets to VPS
+        run: |
+          set -e
+          if [ -f infra/prometheus/targets.yml ]; then
+            scp -q infra/prometheus/targets.yml "$VPS_USER@$VPS_HOST":/tmp/targets.yml
+            ssh "$VPS_USER@$VPS_HOST" "sudo cp /tmp/targets.yml /etc/prometheus/targets.yml 2>/dev/null && sudo rm -f /tmp/targets.yml && sudo systemctl reload prometheus 2>/dev/null || true"
+            echo "✅ prometheus targets synced"
+          else
+            echo "no infra/prometheus/targets.yml"
+          fi
+
      - name: Reload Caddy
        run: |
          ssh "$VPS_USER@$VPS_HOST" "sudo systemctl reload caddy || sudo systemctl restart caddy"
@@ -71,4 +98,4 @@ jobs:
              000|502|503|504) echo "::error::$u bad status $code"; exit 1 ;;
            esac
          done
-          echo "✅ All routes reachable"
+          echo "✅ All routes reachable"
@@ -12,11 +12,10 @@ Reverse-proxy & infrastructure config for [orangevps](https://asepharyana.my.id)
 |------|-----------|-------------|
 | [`asepharyana/hub`](https://github.com/asepharyana/hub) | Portfolio SPA (Next.js, port 4003, dashboard) | `hub` |
 | [`asepharyana/scraper`](https://github.com/asepharyana/scraper) | Rust/Axum scraper API (port 4091) | `scraper` |
-| [`asepharyana/tools`](https://github.com/asepharyana/tools) | Tools: Rust gateway/workers + Next.js frontend (3500/3501) | `tools-gateway`, `tools-workers`, `tools-frontend` |
 | [`asepharyana/llm-api`](https://github.com/asepharyana/llm-api) | Rust LLM API (llama.cpp, port 8080) | `llm-api` |

 Tiap repo punya `flake.nix` + `.github/workflows/deploy.yml` sendiri:
-`nix build .#<pkg>` → `nix copy ssh://` → `nix-env --profile` → `systemctl restart`.
+`nix build .#<pkg>` → **push ke Attic binary cache** (`attic.asepharyana.my.id/asepharyana`) → VPS substitute via `nix-store --realise` → `nix-env --profile` → `systemctl restart`.
 Push ke `main` (atau `workflow_dispatch`) langsung deploy; tidak ada lagi pointer submodule.

 ## Infra di Repo Ini
@@ -4,7 +4,7 @@ Panduan untuk menambahkan service baru di ekosistem `asepharyana/infra` (2026-08

 ## Prinsip

- **Aplikasi hidup di repo sendiri** (`hub`, `scraper`, `tools`, `llm-api`) dengan
+- **Aplikasi hidup di repo sendiri** (`hub`, `scraper`, `llm-api`) dengan
  `flake.nix` + `.github/workflows/deploy.yml` mandiri. Repo infra TIDAK berisi kode app.
 - Repo infra (`asepharyana/infra`) hanya mengatur **reverse proxy & config VPS**.

@@ -51,18 +51,14 @@ imphnen.asepharyana.my.id {
 	import proxy 4009
 }

+gmw.asepharyana.my.id {
+	import proxy 4009
+}
+
 scraper.asepharyana.my.id, api.asepharyana.my.id, scraper.asepharyana.web.id, api.asepharyana.web.id {
 	import proxy 4091
 }

-ai.asepharyana.my.id, ai.asepharyana.web.id {
-	import proxy 4010
-}
-
-tools.asepharyana.my.id, tools.asepharyana.web.id {
-	import proxy 4007
-}
-
 9router.asepharyana.my.id {
 	# LLM streaming: 9router combo models punya TTFT sampe 30-40s (deepseek,
 	# fallback chain). Default (proxy) response_header_timeout 30s / read 60s
@@ -84,7 +80,7 @@ tools.asepharyana.my.id, tools.asepharyana.web.id {
 }

 pr-agent.asepharyana.my.id {
-	import proxy 4002
+	import proxy 4023
 }

 lidm.asepharyana.my.id {
@@ -152,7 +148,34 @@ wiki.asepharyana.my.id {
 		reverse_proxy 127.0.0.1:4020
 	}
 }
+mcpedia.asepharyana.my.id {
+	import proxy 4016
+
+	# Alias mcpedia -> wiki (same app): observability/trpc/hooks/health routes.
+	handle /dashboard {
+		reverse_proxy 127.0.0.1:4020
+	}
+	handle /metrics {
+		reverse_proxy 127.0.0.1:4020
+	}
+	handle /trpc/* {
+		reverse_proxy 127.0.0.1:4020
+	}
+	handle /hooks/* {
+		reverse_proxy 127.0.0.1:4020
+	}
+	handle /health {
+		reverse_proxy 127.0.0.1:4020
+	}
+}

 mcp.asepharyana.my.id {
 	import proxy 4021
-}
+}
+hamc.asepharyana.my.id {
+	import proxy 4024
+}
+
+dev.asepharyana.my.id {
+	import proxy 3000
+}
@@ -11,7 +11,3 @@
  labels:
    service: gmw-backend
    __metrics_path__: /api/metrics
-# llm-api expose /metrics sejak 2026-08-03 (feat metrics):
- targets: ['127.0.0.1:4010']
-  labels:
-    service: llm-api
@@ -0,0 +1,22 @@
+[Unit]
+Description=Hermy HQ — Shorts render worker (Remotion + TTS)
+After=network.target postgresql.service hermyhq.service
+Wants=hermyhq.service
+
+[Service]
+Type=simple
+User=code
+Group=code
+WorkingDirectory=/opt/hermyhq-renderer
+EnvironmentFile=/etc/hermyhq/hermyhq.env
+Environment=RENDER_DIR=/data/renders
+Environment=RENDER_SCRATCH=/tmp/hamc-render
+Environment=REMOTION_ROOT=/opt/hermyhq-renderer/src/remotion
+Environment=CHROME_BIN=/usr/bin/google-chrome-stable
+Environment=EDGE_TTS_BIN=/home/code/.hermes/hermes-agent/venv/bin/edge-tts
+ExecStart=/usr/bin/node /opt/hermyhq-renderer/scripts/render-worker.mjs
+Restart=on-failure
+RestartSec=10
+
+[Install]
+WantedBy=multi-user.target
@@ -0,0 +1,18 @@
+[Unit]
+Description=Hermy HQ — Hermes mission control dashboard (Next.js 16 standalone)
+After=network.target postgresql.service
+
+[Service]
+Type=simple
+User=code
+Group=code
+Environment=NODE_ENV=production
+Environment=HOSTNAME=127.0.0.1
+Environment=PORT=4024
+EnvironmentFile=/etc/hermyhq/hermyhq.env
+ExecStart=/nix/var/nix/profiles/hermyhq/bin/hermyhq
+Restart=on-failure
+RestartSec=5
+
+[Install]
+WantedBy=multi-user.target
Author	SHA1	Message	Date
asepharyanaandClaude Opus 5	21743cedfa	chore(infra): remove llm-api service (retired from prod) Deploy Infra Config / deploy-infra (push) Canceled after 0s Details Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-25 18:53:29 +07:00
MythEclipse	63c3b87f39	fix(hamc): hermyhq port 4023→4024, add hermyhq-render worker unit Deploy Infra Config / deploy-infra (push) Canceled after 0s Details - hermyhq.service: PORT 4024 (4023 occupied by pr-agent-bun) - Caddyfile.prod: hamc.* proxy → 4024 - hermyhq-render.service: new systemd unit for shorts render worker (Remotion + edge-tts, runs /opt/hermyhq-renderer)	2026-09-22 20:16:33 +07:00
asepharyana	c5c0a961b0	chore(caddy): route pr-agent to Bun server (4023) Deploy Infra Config / deploy-infra (push) Canceled after 0s Details Cut-over pr-agent.asepharyana.my.id from the Python pr_agent server (4002) to the Bun/TypeScript re-implementation (4023, pr-agent-bun.service). Verified: webhook -> review -> LLM claude-opus-5 -> persistent comment published (PR #19 nextjs-template).	2026-09-21 13:38:04 +07:00
asepharyana	c25c645818	feat(dev): add dev.asepharyana.my.id Caddy route → 127.0.0.1:3000 Deploy Infra Config / deploy-infra (push) Canceled after 0s Details Ephemeral dev preview of the Next.js template — no systemd service; dev server runs on-demand (bun run dev), Caddy routes :3000, 502s when stopped.	2026-09-20 21:48:22 +07:00
MythEclipse	9e6078af9b	feat(infra): add hamc.asepharyana.my.id (hermyhq) Caddy routing + systemd unit Deploy Infra Config / deploy-infra (push) Canceled after 0s Details - Caddy: hamc.asepharyana.my.id -> 127.0.0.1:4023 (import proxy) - systemd: hermyhq.service (Next.js 16 standalone, profile /nix/var/nix/profiles/hermyhq, env /etc/hermyhq/hermyhq.env)	2026-09-20 15:51:34 +07:00
asepharyana	bda852f287	feat(infra): add mcpedia.asepharyana.my.id Caddy block -> local mcpedia (wiki alias) Deploy Infra Config / deploy-infra (push) Canceled after 0s Details	2026-09-02 12:31:18 +07:00
asepharyana	754bc2876f	fix(caddy): serve gmw.asepharyana.my.id → :4009 (was 525) Deploy Infra Config / deploy-infra (push) Canceled after 0s Details GMW dashboard SSL was broken (Cloudflare 525) because gmw was not in the Caddyfile — no Let's Encrypt cert / TLS origin for the hostname. Add gmw.asepharyana.my.id block proxying to gmw-proxy nginx (4009), same as imphnen. Public dashboard now reachable (200, valid cert).	2026-09-01 20:19:00 +07:00
asepharyana	b37daa52b0	chore(infra): remove tools app — Caddy route, README, add-new-app docs Deploy Infra Config / deploy-infra (push) Canceled after 0s Details Project tools dihapus (2026-08-28): GitHub repo + local source + VPS deploy (3 systemd units + Nix profiles) sudah dihapus. Caddy route tools.asepharyana.my.id removed; README + docs updated. ADR 0003 dibiarkan sebagai catatan historis rename.	2026-08-27 21:17:05 +07:00
asepharyana	56cee21e66	ci(infra): sync full infra config (caddy + systemd drop-ins + prometheus), rename workflow to infra-deploy	2026-08-27 20:43:06 +07:00